COLLECTIVE BARGAINING, STRIKES AND LOCKOUTS
UNDER THE LABOUR RELATIONS ACT, 1995
Prepared for the Ontario Ministry of Labour, to support the Changing
Workplaces Review of 2015
Statement of Work # 10
Timothy J. Bartkiw, LL.M., Ph.D.
November 30, 2015
Timothy Bartkiw is an Associate Professor in the Ted Rogers School of Management,
Ryerson University, where he teaches and researches in the fields of industrial
relations, and labour and employment law and policy.
The author wishes to thank Josh Mandryk for his outstanding research assistance
during the preparation of this Report.
This paper represents the views of the author and not necessarily those of the
Ministry of Labour.
© Queen’s Printer for Ontario, 2016
2
Collective Bargaining, Strikes and Lockouts
TABLE OF CONTENTS
Executive Summary ..................................................................................................... 3
Background .................................................................................................................. 7
Research Findings & Analysis ..................................................................................... 8
Conclusion and Options ........................................................................................... 78
References................................................................................................................. 81
Appendix 1 – Additional Graphs of Work Stoppages Data…………………………......90
Appendix 2 - Summary of Research Design in Quantitive Studies of the Effects of
Specific Components of Labour Policy ….……………………………………
92
Appendix 3 - Provisions in Manitoba Labour Relations Act pertaining to Subsequent
Interest Arbitration Mechanism…………………………………………………………….94
3
Collective Bargaining, Strikes and Lockouts
Executive Summary
This goal of this project was to review statistical trends in Ontario work stoppages, and
to summarize and assess the relevant academic literature examining the functionality of
LRA provisions that are the most salient in respect of the collective bargaining process,
and the operation and duration of strikes and lockouts. These provisions include the
duty to bargain in good faith; the legislative requirements for a legal work stoppage;
alternative responses to labour disruptions (such as subsequent interest arbitration);
issues related to picketing; the reinstatement of employees following a work stoppage;
mandatory and permissive provisions in collective agreements; and the grievance
arbitration system.
Analysis suggests that there has been a general decline in union wage premiums in
Canada in recent decades, suggesting a loss in union bargaining power and/or shift
towards non-wage priorities. On the other hand, small union wage premiums continue
to exist, and these relative advantages are larger in cases where youth, and/or workers
in non-standard employment forms, are able to access unionization. Another indicator
of the functioning of the collective bargaining regime overall, trends in work stoppages,
were also examined. Analysis suggests that for the past few decades, there has been a
steady decline in the occurrence of work stoppages, which is primarily a decline in
strikes, especially in the private sector. An underlying trend is a lack of propensity to
strike in small bargaining units, increasingly common in the modern era, and the already
low propensity to strike amongst small units seems to have been exacerbated over
time. On the other hand, within the increasingly small set of strikes that do occur,
evidence suggests that average strike duration has been growing and that duration
growth is even more pronounced in lockouts and small-unit strikes. A simple startingpoint explanation for the dual pattern of declining strike incidence and duration growth,
on which further explanation can be built, is that there has been a general decline in the
necessity, willingness, or capacity of Ontario workers to strike, in all but cases of
comparably much “larger” conflict or zones of disagreement.
4
Collective Bargaining, Strikes and Lockouts
The subsequent literature review probes for more specific explanations for these
patterns. One body of literature is reviewed that provides potential explanations for
these patterns based on notions of how the instrumentality of strikes is affected by
factors shaping socio-economic context. These factors include: heightened capital
mobility and globalization of production; cultural and ideological forces; shifts in
managerial strategies and capacities; and organizational restructuring. There seems to
be both theoretical and empirical support for the notion that various contextual factors
have produced a general decline in the instrumentality of strikes, which helps explain
the long run decline in strike incidence that we have observed in recent decades.
Another body of literature uses quantitative methods to try to isolate the effects of
specific components of the legal regime governing collective bargaining on work
stoppage patterns, and in some cases also on outcomes such as wages, employment
and investment. These studies have produced a large volume of estimates, many of
which conflict with one another, about the correlation between the presence of specific
labour policy “variables” and the various phenomena being studied. This report
summarizes the results of these studies in detail for the CWR to consider. However,
this report also outlines a number of grounds for the recommendation of exercising
great caution in reliance upon the outcomes of these studies in making policy decisions
about specific aspects of labour law. Some of these concerns pertain to specific studies,
while certain concerns are somewhat common across the body of studies.
Literature examining the regulation of picketing is also examined. This literature
highlights the historical roots of modern picketing regulation, often from a critical
perspective emphasizing the role of judicial ideology in development of legal doctrine.
Other literature examines the 2002 decision of the Supreme Court in RWDSU, Local
558 v. Pepsi Cola Beverages (West) Ltd, the historical events that led to this decision,
and its effects on subsequent development of the common law pertaining to picketing.
Other literature highlights the role of other restrictions on forms of collective action
remaining within the law.
Given empirical patterns of increased strike duration, another section of this report
discusses the possibility of a “subsequent interest arbitration” statutory provision that
5
Collective Bargaining, Strikes and Lockouts
could potentially be adopted to create an alternative policy response to (lengthy) work
stoppages. This report also provides a summary of the subsequent interest arbitration
mechanism that was adopted into Manitoba labour law in 2000, which gives one party
the option of requesting arbitration, subject to various conditions precedent. The report
discusses the minimal existing feedback on the functioning of this mechanism, and also
points out that it has not been invoked very often in Manitoba since 2000, suggesting at
a minimum the lack of any “floodgate” into arbitration in recent years prompted by
adopting this additional mechanism.
This report then provides a review of literature pertaining to the legal Duty to Bargain in
Good Faith. The literature identifies the historical connection with U.S. law, as well as
deviation from U.S. doctrine. It also critically examines various limits of the Duty, and
suggests that its interpretation has enabled the imposition of significant managerial
prerogatives into our labour law over time. Further, heavy reliance upon subjective
intention weakens enforcement of the Duty. A general principle drawn from the
literature is that the Duty is not generally conceived as contributing primarily to the
power-balancing effects of the regime overall, aside from guarding the transition from
individual to collective bargaining per se, implicitly assuming that other aspects of labour
law sufficiently address the balance of power.
The report also examines some of the recent literature about the functioning of the
grievance arbitration system with respect to delay, costs, legalism, and arbitral
jurisdiction. Studies unanimously show a pattern of continued growth, over several
decades, in measures of average duration of time spent on grievances that proceed
through to the end of arbitration. Studies also seek to deconstruct these patterns of
delay, to understand the contributing effects of different factors, at different stages in the
grievance arbitration process overall. Other literature discusses the historical
disappearance of an ethic of “proportionality” in legal process within arbitration, and
argues that any fix to the system must involve a resurgence of this ethic.
Lastly, it is suggested that labour policy overall cannot avoid taking into account the
continued erosion in underlying worker collective bargaining power, one way or another.
It is possible that policy-makers may continue to refuse to adapt the collective
6
Collective Bargaining, Strikes and Lockouts
bargaining regime so as to potentially counter the corrosive effects on worker collective
bargaining power produced by various contextual shifts. Perhaps this may be out of a
failure to value worker collective bargaining power. Alternatively, it could be due to a
preference for other types of instruments, beyond the collective bargaining regime, as
potentially targeted responses to specific forms of work/employment precarity. Leaving
aside the highly contested desirability of such a choice, effective labour policy overall
demands an honest recognition of the decline in worker collective bargaining power and
its consequences, in order to either address it directly, or to assess the sufficiency and
effectiveness of all other forms of labour policy responses to increased inequality and
work/employment precarity, absent policy reform ameliorating the underlying decline in
worker collective bargaining power.
7
Collective Bargaining, Strikes and Lockouts
Background
The Labour Relations Act, 1995 (“LRA”) contains provisions pertaining to the
negotiation, content and operation of collective agreements. The design features of
these legislative provisions may have impacts on the establishment and subsequent
security of unions, as well as on outcomes such as strike incidence, strike duration, and
wages. These issues are of great concern to the parties directly involved in collective
bargaining, but also to the public.
8
Collective Bargaining, Strikes and Lockouts
Research Findings & Analysis
1) Introduction:
This report provides an examination of certain data and literature pertaining to the
functionality of the collective bargaining regime. Section 2 of the report provides an
examination of certain key empirical trends. Recent trends in the size of union wage
premiums are reviewed, as are trends in work stoppages. Section 3 provides a review
of literature on the functionality of the legal regime governing collective bargaining and
work stoppages. This section includes various subsections containing reviews of
literature concerning: strikes and lockouts as instruments in socio-economic context; the
effects of various components of the collective bargaining legal regime on work
stoppages and related outcomes; the regulation of picketing activities; subsequent
interest arbitration as an alternative response to work stoppages; the duty to bargain in
good faith; and the grievance arbitration system.
2) The “big picture”: some key empirical trends
This section reviews certain key empirical indicators of the functioning of the collective
bargaining regime in recent decades. This includes analysis of recent trends in the size
of the union wage “premium,” and in work stoppage activity.
a) Some recent trends in measures of the union wage “premium”
Historically, evidence suggests that unionized workers in Canada have tended to have
somewhat better conditions of employment than comparable non-unionized workers,
and thus it has been possible to measure the union wage “premium”, or the gap in
compensation levels of comparable unionized and non-unionized workers, ceteris
peribus. Recent analysis suggests that union wage premiums have been declining in
recent decades, suggesting a decline in union bargaining power, or sometimes called a
decline in so-called “wage-setting” power, and/or a shift in union focus on non-wage
outcomes. Fang and Verma (2002) found that the union wage gap towards the end of
1990’s declined to an overall average of approximately 7.7%, suggesting a gradual
9
Collective Bargaining, Strikes and Lockouts
narrowing of the wage gap over the previous two decades from earlier estimates of
about 15% (Benjamin et al., 2012, at 466) The gap was comparably smaller in large
firms, and also varied by a number of factors, such as industry, occupation, and region.
Similar effects have been observed in the U.S. in recent decades.
Subsequent analysis suggests that the average union premium continued its steady
decline into the mid-2000s in Canada. After confirming the continued decline in the
Canadian union wage premiums from 2001 to 2006, Walsworth and Long (2012) also
examined trends in the effect of unionization on employment, and the relationship
between these trends. Their analysis suggests that there has been a significant decline
in the effect of unionization on employment, and that there is evidence to suggest that
this is somewhat explained by the decline in union premiums. They also found that
while employment suppression from unionization continued in large manufacturing
firms, in smaller firms there was an opposite effect. That is, amongst smaller firms,
union premiums were somewhat negative (-1.9%), and unionization was associated
with employment growth. This pattern suggesting that unions were trading wage
advantages for jobs generally held for the services sector as well.
Verma, Reitz and Banerjee (2015) examine the effects of unionization on the labour
market integration of newly arrived immigrants in Canada. Compared to white members
of their immigrant cohort, non-white recent immigrants experience slower access to
unionized jobs, and a smaller effect of unionization on their earnings. Their study
provides evidence that since 1993, unionization has not contributed to closing the
specific earnings gap between white and non-white immigrants.
The most recent available analysis of union wage premium dynamics is Gomez and
Lamb (2015). Their analysis shows that in aggregate terms, union wage premiums
underwent a slow decline during the period 2000-2012. While women overall
experienced a larger premium than men during these years, women’s union wage
premiums declined from approximately 11% to 7.4%, while male union wage premiums
declined from approximately 6.8 to 5.3% respectively. Gomez and Lamb also assessed
differences in union wage premiums amongst workers with full-time permanent
positions, and amongst workers who lacked full-time permanent status. The union
10
Collective Bargaining, Strikes and Lockouts
wage premium amongst workers lacking full-time permanent status was much larger
than amongst workers that had this employment status (17-24% premiums compared to
1-2% premiums respectively). This suggests that workers in non-standard employment
forms have much to gain from unionization, even if non-standard employment forms are
maintained, where such workers are able to access unionization. Also, the interaction
of gender in determining or mediating the size of union premiums has diminished
substantially. While women had generally higher premiums in the full-time permanent
category and men higher premiums in the non-standard or “precarious” category, by
2012, these premiums seem to have converged, eliminating gender differences in the
size of premiums within both employment form categories. Similarly, Gomez and Lamb
(2015) also found that among young workers (aged 15-29) the union premium rate is
comparably much higher than the rate for all employees. Further, amongst young
workers lacking full-time permanent status, union premium rates in this category were
much higher than the average premium for youth. This suggests that young workers in
non-standard employment forms have a large potential gain from unionization, even if
non-standard employment forms are maintained, where such young workers are able to
access unionization.
Overall then, there has been a general decline in the magnitude of union wage
premiums in recent decades, suggesting a loss in union “wage-setting power” over time.
On the other hand, some union premiums continue to exist, and appear largest in
contexts involving non-standard employment forms and/or youth employment. The
general decline in union premiums may be explained by a number of developments in
socio-economic context in recent decades, including heightened international product
market competition, heightened capital mobility, and patterns in industrial restructuring.
These developments may be understood as having narrowed the space for wage gains
under unionization due to increasing constraints of competition, and/or as having
caused shifts in the balance of power in the collective bargaining process, eroding the
union advantage.1 The comparably higher union premiums for non-standard
1
The comparably higher union premiums that exist for non-standard employment and youth largely reflects, on
the one hand the comparably greater precarity experienced by these workers, in non-union contexts, while on the
other suggesting the potentiality of unionization, where it may emerge, to counter this precarity.
11
Collective Bargaining, Strikes and Lockouts
employment and youth largely reflects, on the one hand the comparably greater
precarity experienced by these workers, in non-union contexts, while on the other
suggesting the potentiality of unionization, where it may emerge, to counter this
precarity. The next section examines work stoppages patterns in the collective
bargaining process.
b) Patterns in work stoppage activity
This section provides a basic overview of some of the relevant historical patterns in
work stoppages within the collective bargaining system over time. It has been argued a
great many times that the strike is the ultimate source of worker bargaining power in the
collective bargaining system in Canada, and elsewhere.2 Therefore, an understanding
of strike dynamics may provide some important insights into the functioning of the
collective bargaining system overall.
Data in this section was provided by ESDC. Unless otherwise stated, aggregated data
is based on aggregate activity in bargaining units of all sizes,3 during the time period
1980-2014.
To begin with, a very simple observation can be made, which is that there has been a
fairly clear trend of a long term decline in work stoppages in Ontario, and that this has
been primarily the product, in an accounting sense, of a long term decline in strike
activity. Figure 1 provides a graph of the aggregate number of work stoppages, strikes
and lockouts in Ontario4. As can be seen in Figure 1, there remains significant
fluctuation in strike activity over time, although the long term decline is also clear.
Lockouts have generally been far less frequent, and there is less of an obvious pattern
2
While it may be that, symetrically, the lockout is the ultimate source of the employer’s bargaining power, this
seems less collectively accepted.
3
There is one small caveat to this. ESDC records data for all work stoppages totaling ten or more person days lost.
This could be either one worker for ten days, or ten workers for one day, etc. So, there is a small chance that this
inclusion threshold may filter out some work stoppage activity in very small units. All underlying data obtained
from ESDC is available from the author upon request.
4
ESDC categorizes work stoppages as taking place in Ontario when Ontario is the relevant legal jurisdiction. Work
stoppages physically occurring within Ontario, but which fall under the jurisdiction of federal labour laws, are not
included in this data.
12
Collective Bargaining, Strikes and Lockouts
over time pertaining to lockout activity.5 This trend has held generally across Canada
and beyond (Godard 2011). Figure 2 provides a graph of the aggregate number of
work stoppages, strikes and lockouts in all Canadian jurisdictions, for all unit sizes. The
fact that there has been a similar aggregate trend across Canada is fairly clear.
This picture of a major decline in strike activity also shows through when measures of
the volume of strike activity are used other than the aggregate number of
Figure 1: Aggregate Annual Number of Work Stoppages, Strikes, and Lockouts in
Ontario, 1980-2014
5
As can be seen from the graph, there was a small “spike” of strike activity in 2012, and subsequent graphs show a
corresponding “plunge” in strike duration in that year. Both of these measures seem influenced by the occurrence
of a single one-day teacher strike event that took place across Ontario in that year, which HRSDC recorded as 75
separate strike events.
13
Collective Bargaining, Strikes and Lockouts
Figure 2: Aggregate Annual Number of Work Stoppages, Strikes, and Lockouts in
Canada, 1980-2014
Figure 3 – Aggregate Annual Number of Workers Involved in Work Stoppages,
Strikes and Lockouts in Ontario, 1980-2014
14
Collective Bargaining, Strikes and Lockouts
strike events. For example, Figure 3 provides a graph of the number of workers
involved in stoppages, strikes and lockouts in Ontario. Although there is certain
variation, and a surge in the number of workers involved in 1996-97, a long term decline
trend is apparent. A comparable pattern in the numbers of workers involved in
stoppages, strikes and lockouts also holds for across Canada, and a graph illustrating
this is in Appendix 1.
The data also show that this long term decline in strike activity is also not merely a
simple product of a decline in the number of unionized workers in the province. Figure
4 provides a graph of the ratio of the number of workers involved in stoppage, strike and
lockout activity to aggregate unionized employees. As can be seen in Figure 4, there is
an observable long term decline in the proportion of unionized workers engaged in work
stoppage events in Ontario. Again, this Ontario experience is reflected in a similar
larger pattern in the data for Canada overall, and a graph of this ratio for all of Canada
is provided in Appendix 1.
Figure 4: Aggregate Number of Workers Involved in Work Stoppages, Strikes and
Lockouts as Ratio of Total Unionized Employees (Ontario) (1980-2014)
15
Collective Bargaining, Strikes and Lockouts
So, thus far, it appears that there has been a long term decline in stoppages activity,
captured in different indicators, in both Ontario and all of Canada, and that in an
accounting sense, this is the product of a long term decline in strike activity. A further
important insight provided by the data relates to the existence of different trends in work
stoppages across the different sectors. Figure 5 provides a graph of the number of
strikes in Ontario that occurred in the public and private sectors, and Figure 6 provides
a graph of the number of workers involved in strikes in Ontario, also by sector. 6 The
immediate insight from these two graphs is that while there is not a terribly clear longterm pattern in the public sector, the bulk of the decline in strikes appears to have
occurred in the private sector. Further, this decline in private sector strike activity does
not appear to be a mere product of a decline in private sector unionization, since the
magnitude of the decline in strike activity seems fairly disproportionate to the rate of
Figure 5: Aggregate Number of Strikes in Ontario by Sector (1980-2014)
6
Data is allocated to the different sectors by ESDC, according to official definitions of these used for statistical
purposes. The precise definitions used to assign data to the public sector is available at:
http://www.labour.gc.ca/eng/resources/info/datas/wages/technical_notes.shtml
16
Collective Bargaining, Strikes and Lockouts
Figure 6: Number of Workers Involved in Strikes in Ontario by Sector (1980-2014)
Figure 7 – Number of Private Sector Strikes by Small Bargaining Units (1-49
Employee units, and 1-99 Employee units) Compared to the Total Number of
Private Sector Strikes in Ontario (1980-2014)
17
Collective Bargaining, Strikes and Lockouts
change in union density, and/or unionized employment levels, in the private sector.7
Thus, there seems to have occurred a significant overall decline in the necessity,
willingness, or capability, of Ontario private sector workers to strike in recent decades.
Further insights pertaining to this pattern of strike decline may also be available by
looking at differences in the distribution of strikes by bargaining unit size. Figure 7
provides a graph comparing the total number of private sector strikes to those in “small”
bargaining units, using both 1-49 employees and or 1-99 employees as definitions for
“small” units. It appears that strikes have been declining across all bargaining unit size
categories. In certain research that involved econometric modelling of strike behaviour,
discussed later in this report, the role of bargaining unit size was found to be statistically
significant and suggested that smaller bargaining units, all other factors being equal,
have a lower propensity to engage in strike action.8 However, more is seemingly at play
here than a mere decline in average unit sizes, given that the number of strikes within
the small units categories also declined over time, suggesting that the lack of strike
propensity in small units may have been exacerbated over time.9
The “size” of strike activity may also be measured with respect to its time duration.
Figure 8 provides a graph of the average duration of work stoppages, strikes and
lockouts in Ontario, in units of all sizes, and a similar graph for all of Canada is included
in Appendix 1. A few points can be gleaned from this graph. It seems that the average
duration of lockouts has varied substantially over time, to a greater degree than strike
duration, although this is potentially related to the fact that there are far fewer lockouts
than strikes, making the average of this category more prone to wider variation. In more
recent years (since 2000) there is some evidence of a trend towards longer average
lockout duration, although variation remains strong. Further, the degree of variation in
7
For example, Galarneau and Sohn (2013) estimate that Canadian private sector union density declined from
18.4% in 1999 to 16.4% in 2012, while private sector employment growth would also have offset some of the
density decline effect on the total number of unionized workers in the private sector. The proportional decline in
private sector strikes (and workers involved) appears much greater than the decline in private sector unionization.
8
See Hebdon, Hyatt and Mazerolle (1999); Campolieti, Hebdon and Hyatt (2005); and Campolieti, Hebdon and
Dachis (2014).
9
Data on the underlying distribution of unionized employment by bargaining unit size, for comparison purposes,
was not available.
18
Collective Bargaining, Strikes and Lockouts
average strike duration also seems to have widened in recent years. Based on the
graph alone, there appears to be some evidence of an increase in strike duration,
although the variation makes this somewhat difficult to confirm. In particular, the year
2012 seems to be somewhat of an “outlier” compared to the years close to it, in that it
has quite a low measure for average duration. However, this statistic seems very
heavily influenced by the existence of elementary school teacher bargaining in Ontario
in that year, within which there was a province-wide, one-day strike that occurred. In
response to this single coordinated event, ESDC recorded 75 separate one-day strike
incidences, in 75 regions. This greatly inflated the strike/stoppage incidence
measurement and reduced the average duration measurement, for 2012.10 Indeed,
some academic analysis of strike/stoppage duration takes the position that all such
activity of a fixed, pre-planned duration ought to be excluded from analysis of duration,
Figure 8 – Average Duration of Work Stoppages, Strikes and Lockouts, Ontario
(1980-2014)
10
All of these one-day teachers strikes took place in the public sector. I have confirmed with ESDC that the data
for 2012 for the private sector were not affected by these teacher strikes. So, 2012 remains an outlier year for the
Ontario private sector, with comparably high strike incidence and low average strike duration, coincidentally
occurring in the same year as these 75 one-day teacher strikes.
19
Collective Bargaining, Strikes and Lockouts
Figure 9: Average Duration of Work Stoppages, Strikes and Lockouts in Ontario,
in private sector (1980-2014)
because very different phenomena are being measured (Finley, 2010). So, if we were
to take into account the fact that the average duration for 2012 is somewhat “artificially”
suppressed, then the trend towards longer duration becomes somewhat clearer. In
terms of stoppages overall, there seems even clearer evidence of growth in average
duration in recent years. These comments would appear to apply similarly to activity in
Canada overall, as illustrated by another graph in Appendix 1. 11
Finally, Figure 10 provides a graph of average strike duration in the private sector for
bargaining units of different size ranges. Interestingly, the graph seems to provide
evidence of the existence of comparably longer strikes in smaller bargaining units than
in larger ones. This may be related to patterns of industrial restructuring and
organizational “fissuring,” and a resulting decline in the instrumentality of strikes,
discussed further in the next section.
11
Gunderson at al (2009, at 344) provide a similar analysis, although based on comparisons with earlier historical
experience, and based on stoppages data overall.
20
Collective Bargaining, Strikes and Lockouts
Figure 10: Average Duration of Private Sector Strikes in Small Bargaining Units
(1-49 Employee units, and 1-99 Employee units) Compared to Average Duration of
All Bargaining Units (1980-2014)
Overall, the data suggests that for the past few decades, there have been fewer work
stoppages overall, and in an accounting sense this is mostly because there are fewer
strikes, especially in the private sector. An underlying trend is a continued decline in
strikes within smaller units, and past studies had suggested that small units already had
comparably lower strike propensity. In contrast, within the smaller set of existing
strikes, evidence suggests that average duration has been growing, and duration
growth is even more pronounced if we look at stoppages overall, including lockouts, and
is also more pronounced in smaller workplaces.
There is no one single definitive explanations for these combined trends of decreased
strike incidence and increased average duration. The “Annis Report” came to similar
conclusions about patterns of work stoppages incidence and duration in the federal
jurisdiction, but stated that research had not yet provided a conclusive interpretation for
increased duration (Annis, 2008). This view seems at least partly due to the
21
Collective Bargaining, Strikes and Lockouts
acknowledgement of alternative theoretical approaches in examining strike dynamics,
discussed more fully below. However, a fairly simple starting point interpretation, one
that admittedly begs for further probing, is that there has been a general decline in the
necessity, willingness, or capacity of workers to strike, in all but cases of comparably
much “larger” conflict or zones of disagreement. Whether this in turn has been based
upon an increase in the “costliness” of strikes, a decline in their efficacy, or some other
exogenous factors, this simple interpretation at least provides an initial explanation for
both patterns of declining strike incidence, and growth in average duration. The next
section looks more closely at this question of how to interpret these work stoppage
patterns, and to potentially understanding the role of the legal regime governing
collective bargaining, and other contextual factors, in shaping these patterns.
22
Collective Bargaining, Strikes and Lockouts
Section 3 – Review of Academic Literature on the Functionality of the Legal
Regime Concerning Work Stoppages and Collective Bargaining
This section of the paper will review literature pertaining to salient components of the
legal regime governing collective bargaining identified by the CWR for review. This
section reviews literature that attempts to analytically isolate the effects of specific,
separate components of the collective bargaining legal regime identified by the CWR,
but begins with a review of literature examining the strike instrument in socio-economic
context.
A) Strikes and lockouts as instruments in socio-economic context
Up until the mid-late 1980’s, econometric analysis of strike data suggested that there
had long existed a significant positive relationship between strike incidence and the
business cycle, based on the notion that strikes had greater efficacy during peaks than
troughs in the business cycle (Rees, 1954; Ashenfelter and Johnson, 1969; Smith,
1972). This relationship held not only in Canada but in the U.S. as well (Ashenfelter
and Johnson, 1969). Research suggested that the positive relationship between strike
frequency and the business cycle was relatively stronger in North America than in many
western European countries. This was likely because in these countries, the influence
of corporatist systems of centralized bargaining for much of 20th century outweighed or
mitigated the influence of the business cycle, unlike in North America (Brym, 2009).
However, in both Canada and the U.S., beginning in the 1980’s, the influence of the
business cycle on strike incidence seems to have been greatly suppressed. Strike
incidence has steadily declined since then, with cyclicality largely “wrung out” of the
system, displaced by a more generalized downward trend in recent decades (Brym,
2009), confirmed by the data review in the previous section of this report. This suggests
that other contextual factors have had a more overwhelming effect on these dynamics,
by altering the instrumentality of strikes over time.12
12
On the other hand, previous econometric research suggested a different effect of the business cycle on strike
duration. Harrison and Stewart’s (1989) analysis of Canadian strike data from 1946-1983 supports the thesis that
23
Collective Bargaining, Strikes and Lockouts
Before the various contextual factors are reviewed, some discussion of the different
theoretical models concerning strike activity is helpful at this point, since they involve
different understandings about the degree to which contextual factors may “matter” in
determining strike dynamics. There is no single, commonly accepted theoretical
“model” of strike activity that can be turned to in order to determine the effect of
contextual factors on strike dynamics (Kaufman, 1992). A number of contextual factors
that we will review point to a shift in relative bargaining power from unions/workers to
employers. Some theoretical approaches would see these factors therefore as being
relevant in explaining strike dynamics in a fairly direct manner.13 Over time, however,
certain theoretical approaches have taken issue with bargaining-power-based
explanations, reasoning that these factors are better conceived as strictly affecting
wage outcomes, but not strike activity itself (see Gunderson, et al., 2008). In their view,
shifts in relational power, known in advance by rational parties, ought to be conceived
as creating predictable effects on the expected outcomes of bargaining, which is then
factored into parties’ bargaining demands and/or willingness to concede.14 Therefore,
only the outcomes of bargaining shift, the propensity to strike does not. This theoretical
challenge is rooted heavily in the “neoclassical” theoretical approach, which in its default
form involves strict starting-point assumptions about the functioning of a conceptualized
“perfect” market including, but not limited to, “perfect” (so-called) “rationality”, and
perfect information. Under such assumptions, however, it turns out that there are
actually extremely few, if any, explanations for strike activity whatsoever. This is
because in a world in which parties are perfectly “rational” and have a “perfect” grasp of
all information about their own15 (and their bargaining opponent’s) true status,
strike durations were countercyclical, with fairly strong magnitudes and robust effects across different statistical
indicators of the business cycle.
13
See for example early bargaining power based strike models developed in Neil W. Chamberlain and James W.
Kuhn (1965), Collective Bargaining. New York: McGraw-Hill.
14
Commenting on the difficulty of rational-choice model-building in the strike literature, Kennan (1986) referred to
this overall as the “Hick’s paradox” and stated “if one has a theory which predicts when a strike will occur and
what the outcome will be, the parties can agree to this outcome in advance, and so avoid the costs of a strike. If
they do this, the theory ceases to hold.”
15
This of course also ignores the added complexity of the fact that we are dealing in these contexts with attempts
to aggregate preferences across multiple workers in a bargaining unit, and further ignores differences in interests
of different kinds of stakeholders on the “employer” side of bargaining. It should be noted that some
24
Collective Bargaining, Strikes and Lockouts
preferences, and measures of the costs/benefits of negotiated terms, and the costs of
strike action on each, parties would always be incented, and able, to carefully contract
so as to avoid strikes (assuming strikes produce costs), under given constraints. This
neoclassical theoretical counterfactual sheds light on why a large literature is referred to
collectively as “strikes as mistakes,” since they “theoretically” should not occur. This
label discursively privileges the neoclassical theory as a default, contestably. Taken
strictly, the argument that context-based bargaining-power shifts are only capable of
effecting wage outcomes - and not strike behaviour – is a theoretical move that reasserts neoclassical assumptions rejected in other theoretical approaches.
Forced to acknowledge some degree of real-world imperfections in inter alia information
and rationality, the two main models that predominate the economics-based literature
and which still seek to develop a “bounded” rational-choice approach to modelling strike
dynamics are the “joint cost” model and the “asymmetric information” model. Under the
former, the lower the joint costs of the strike, the more likely strikes are to occur. Thus,
factors that increase the joint costs of strikes are likely to reduce strike activity, and vice
versa. Under the asymmetric information model, unions are assumed to have
comparably less information about the true measure of the employer’s willingness to
pay, and the employer faces an incentive to withhold this information. Strikes are then
an instrument for “revealing” the true status and preferences of the employer, by forcing
the employer to “pay” in order to maintain its stated demands.16 Overall, these models
leave comparably more “space” for the interpretation that context-based bargaining
power shifts help explain strike dynamics.
Under other theoretical approaches besides these two economics-based approaches,
explanations for the (non)occurrence of strike activity are based around a core
assumption that strikes are the product not only of imperfections in information and
rationality, but also of highly complicated processes of social and/or psychological
behaviour, social interaction, preference construction, and collective mobilization
(Godard, 1992, 1998). These approaches provide additional scope for interpreting
“neoclassical” analysis (if it is fair to still call it that after the acknowledgement of the violation of most of its initial
postulates) has attempted to “take into account” some of these critiques of this approach.
16
See the review of these models in Gunderson et al. (2008).
25
Collective Bargaining, Strikes and Lockouts
context-based shifts in bargaining power and strike instrumentality as also affecting
strike incidence and/or duration.17
One key contextual factor cited in literature is the significant increase in recent decades
in international product market competition, trade liberalization, and capital mobility
(Brym, 2009). If strikes are costly, and potential gains for labour under collective
bargaining are narrowed in certain contexts by international product market competition,
then net gains from strike activity are biased downwards, reducing their
instrumentality.18 Further, particular employers are more direct “beneficiaries” of new
modes of mobility. This can be expressed in different ways. In some cases, certain
employers have an expanded option of “exit”, potentially involving the closing or
relocation of an operation to a jurisdiction with lower labour costs and/or standards.
This obviously affects relational bargaining power, and may correspondingly reduce the
efficacy of the strike instrument. Indeed, an even more direct effect may also be
operating such that, as a result of the new context, certain employers may be
comparably able to avoid the effects of strikes by being able to reorganize the
productive activities of the enterprise by utilizing productive capacities outside the
jurisdiction in which the labour dispute is taking place (Brym, 2009; Shniad, 2010;
Peters, 2010). Such reorganization of activities may take place during and in response
to a strike itself, while certain reorganization efforts may also be carried out ex ante so
as to signal an increased employer readiness to resist strike action. Strategies of
reorganization of work activities outside the jurisdiction have been cited as relevant in
literature examining several recent large-scale, private sector strikes. These include the
2005 Telus strike (Schniad, 2010) the 2009 Vale-Inco strike in Sudbury, Ontario (Peters,
2010); the 2012 Caterpillar strike in London, Ontario (Olive, 2012; Brym et al., 2013)
and the Navistar International dispute in 2011 (Olive, 2012).
17
Brym (2009) cites this challenge to his review of socio-economic factors as really being a model of wage effects,
without providing much of a defensive response, in a footnote. See Brym, 2009, at 73, note 3. Nevertheless, it is
fairly clear that Brym is theorizing that these socio-economic phenomena have had effects on both relational
power (and hence wages) as well as strike efficacy and frequency. See also John Godard (2011).
18
In interpreting the effect of liberalized trade on strike incidence, it is possible to use the “joint cost” and
“asymmetric information” models to provide somewhat similar predicted effect. Such an approach would
seemingly necessitate yet further assumptions including heightened power of consumers, and/or allegedly greater
revelation of the employer’s “ability to pay,” under liberalized trade.
26
Collective Bargaining, Strikes and Lockouts
The notion of increasingly globalized production and heightened capital mobility must be
understood as conceptually distinct from the concept of “foreign ownership” that has
been used as an independent variable in two past econometric studies of strike data.
Based primarily on theoretical models of negotiation behavior and the possibility that
foreign ownership may create greater “informational” challenges and thus higher risks of
“mistakes”, prior studies produced conflicting results. Cousineau et al. (1991) found that
foreign ownership was negatively correlated with strike incidence, prompting them to
interpret this result as suggesting that foreign-owned firms likely develop compensatory
“protocols” addressing informational problems that might otherwise generate strikes.
Using additional control variables omitted in previous studies, such as industry
categories, and bargaining unit size, and different sub-categories of corporate
ownership status, Budd (1994) also examines the effect of foreign ownership on strike
dynamics using Canadian strike data from 1965-1985. Budd (1994) found that there
was no statistically significant effect of foreign ownership on either strike incidence or
duration. Not only is the concept of foreign ownership different from this concept, these
studies were also examining the effect of “foreign ownership” within older data,
pertaining to an earlier era predating the “neoliberal” era of heightened capital mobility.
Generalized welfare state restructuring has also been cited as an important contextual
factor with potential influence on strike efficacy and thus strike activity. Over recent
decades, welfare state shifts have included reductions to income replacement benefits,
and increased restrictions on access to such benefits. Given the (perceived) potential
for loss of short term income security during strike action, and/or some (perceived)
potential for permanent job loss in relation to strike, this adjusts the (perceived) costs
and benefits, and related uncertainty and riskiness of strike action (Brym, 2009). 19
19
A recently examined Service Canada website explains that in general, workers in 2015 Canada that are
affected by any “labour dispute,” including either a strike or lockout, are not eligible for Employment
Insurance. More detail about this restriction on access to EI benefits is available at:
http://www.servicecanada.gc.ca/eng/ei/types/regular.shtml
27
Collective Bargaining, Strikes and Lockouts
Another factor cited in the strike literature that is associated with some other contextual
factors, but may also be potentially understood as a separate independent factor, has
been cultural or ideological shifts over recent decades. In general, the argument is that
the shift to a more neoliberal political economy has coincided with an increasing
embrace of the idea that control over labour costs is a desirable, and priority goal,
making strike activity, and perhaps even unionization itself, increasingly counter to
dominant cultural norms (Brym, 2009).
Another contextual factor is managerial strategies. First, it is possible that there have
been shifts in managerial responses to strike action. Some historical and qualitative
analysis cites shifts away from long-term stable patterns in managerial behaviour
towards much more aggressive responses to strike action (Schniad, 2010, Peters,
2010, Brym et al., 2013). Increasingly aggressive responses to strike activity may
include greater inclination to use replacement workers, or a more intensive use of them;
the transfer and/or contracting out of work; the ex ante development of more
sophisticated business continuity plans in anticipation of strikes; the use of professional
security firms specialized in “controlling” strike/picketing activity;20 and aggressive
litigation strategies seeking to restrict or deter picketing activity. Given an increase in
managerial capacity to undertake these strategies, this may on the one hand operate as
a deterrent to strike activity. On the other hand, it may reinforce some employers’
willingness to either engage in a lockout or to “take a strike,” as a mechanism for
obtaining substantial concessions. The latter view has been provided as an explanation
for the observed pattern of numerous large-scale stoppages of long duration in recent
years (Stanford, 2012; Brym et al., 2013; Oved, 2015). As well, qualitative studies of
20
The argument has been made that picketing workers may experience these services as harassment, and thus
they may undermine support for collective action (Schniad, 2010). Discussions of these services have emerged in
recent policy dialogue (MacDonald, 2015) and in corporate literature on “business continuity plans” during strikes.
For an example of the latter, see the report documenting the development of a sophisticated “business continuity
plan” by one of these specialist firms for Telus, about a year and a half prior to the 2005 Telus strike. This is
available for download at http://www.afimacglobal.com/pmartin/wp-content/uploads/2011/01/GlobalAssurance-x3.pdf.
28
Collective Bargaining, Strikes and Lockouts
certain stoppages have provided evidence that the use of replacement workers tends to
increase strike duration (Singh, Zinni, and Jain, 2005).21
Given the significance of the issue of the employer’s right to use replacement workers in
our regime, there are few studies pertaining to the actual use of replacement workers,
as an employer strategy. One study based on Ontario data in the year 1991 suggested
that replacement workers were used in only 13.8% of work stoppages, and that external
replacements (from outside the bargaining unit) were used in only 12% of all
stoppages.22 In a sample of strike data from the federal jurisdiction from 1991-1999,
used in Jain and Singh (1999), replacement workers were used in approximately 18% of
strikes. While we lack aggregate statistical evidence on employer inclination to use
replacement workers duration strikes, certain case studies of large-scale conflicts
suggest that some employers have begun to use them, contrary to their own past
practices.23
Of course, shifts in employer strategies that may have affected strike dynamics are not
limited to ad hoc responses to strike action. Other ex ante strategies are relevant,
particularly organizational “fissuring” (Weil, 2014) carried out with increased intensity
over the past few decades. David Weil, Director of the Wages and Hours Division of the
U.S. Department of Labor, offers the following visual analogy:
“The large corporation of days of yore came with distinctive borders around its
perimeter, with most employment located inside firm walls. The large business of
today looks more like a small solar system, with a lead firm at its center and
smaller workplaces orbiting around it. Some of those orbiting bodies have their
own small moons moving about them. But as they move farther away from the
21
Of course, the flow of causality is always a theoretical question. One theoretical interpretation of the evidence
pertaining to the relationship between replacement workers and duration is that management hires replacements
based on their advanced knowledge (which must implicitly be assumed in this neoclassical-economic inspired
interpretation) that the strike will be long. The need for more sophisticated analysis of causality to aid in such
theoretical interpretive conundrums is regularly cited as one of the grounds in favour of more intensive qualitative
analysis, such as the research design employed by Singh, Zinni, and Jain (2005) (see Palys, 2003).
22
See Haywood (1992). The data in this paper was further reviewed in Langille (1995) prompting him to suggest
that the restrictions on replacement workers in Bill 40 (passed in 1993) likely had little effect on actual behaviour,
since employers used replacements so rarely before its passage.
23
See Peters (2010) (Re Vale-Inco conflict).
29
Collective Bargaining, Strikes and Lockouts
lead organization, the profit margins they can achieve diminish, with consequent
impacts on their workforces (Weil, 2014 at 43). (emphasis added)
Crucially, Weil’s analysis points to the continued existence of “lead firms,” which exist in
a vast number of industries and are the epicentres of organizational “fissuring.” To
make fissuring feasible and profitable, lead firms must retain certain key organizational
capacities, functions, and responsibilities relating to productive activities that they
coordinate. Over time, developments in information and communications technology,
and new legal and organizational capacities, effectively lowered the “transactions costs”
(eg. monitoring costs, contractual enforcement costs, etc.) involved in having productive
activities performed by workers located (at least in formal terms) “outside” the lead firm.
Lead firms then increase their relative focus on the revenue side of the enterprise,
maintaining responsibility for marketing, branding, and supervision of the larger
contractually-linked network. These new structural arrangements have taken many
forms, including franchising, subcontracting, and supply-chain arrangements.24 In
formal legal terms, the new separate and subordinate organizations typically also
became defined as the separate, sole “employers” of “their” workers. Of course, legal
tests and rules pertaining to de jure “employer” status constrain the degree to which
lead firms can alter employer status through organizational restructuring, and lead firms
also still need to balance the potential gains from this outcome against their need to
maintain a sufficient degree of control over subordinate firms (and/or “their” workers) for
fissuring to be successful, from the perspective of the lead firm. Weil’s analysis
suggests that employer status tests for determining the single “true” employer, weighted
heavily in favour of conceptions of “control” over work activities, have provided ample
scope for lead firms to strike the balance needed for successful fissuring to occur.25
24
While he does not define it as a separate category in its own right, Weil also cites the expansion in temporary
help agency employment as an important example of fissuring and, although its categorization is not precisely
specified, he seems to treat agency-supply relationships as being specific examples of subcontracting and/or
supply-chain structures.
25
Weil’s claim about the relationship between employer-status rules and fissuring is within the U.S. context. The
degree to which the functioning of Canadian employer status rules in labour law may have affected the relative
degree of fissuring in Canada is not known. Although employment law in both countries has much in common in
the approach to allocating employer status, some substantive differences seem to exist. See Sack, Phillips and
Leal-Neri (2011).
30
Collective Bargaining, Strikes and Lockouts
This shift of work/workers “outward” has created various advantages for lead firms.
One key advantage has been that it has enabled lead firms to avoid direct employmentrelated responsibility and liabilities while still obtaining its necessary labour supply.
Arguably, this focus on liability avoidance has been the primary emphasis in literature
discussing these organizational trends, and has also been the primary focus of related
policy dialogue.26 However, a deeper structural effect of fissuring that has been less
emphasized in policy dialogue is the generalized erosion of worker bargaining power
produced by fissuring (Bartkiw, 2015). As Weil notes, fissuring alters the underlying
wage determination process itself. Fissuring enables lead firms to locate work/workers
into “spaces” in which there is comparably more competition in labour supply. This is
partly the result of the fact that in the new “product” market that the subordinated firm
now competes, there may be other competitors (or potential competitors) with respect to
the supply of this service to the lead firm. Heightened competition is also a product of
comparably lower barriers to entry into the narrower scope of business activity
performed by the subordinate firm. Lead firms may capture some of the benefits of this
competition, leaving comparably smaller profit margins for subordinate firms, and
increased competition based on labour costs.
Organizational incentives to fissure, and its consequences, are logically mediated by the
nature of the legal regime governing collective bargaining and the predominant patterns
in bargaining structure that it fosters (Bartkiw, 2015). In the Canadian private sector
context, a highly decentralized bargaining structure has been imposed by the legal
regime, given the default rule that workers within a defined bargaining unit must share
inter alia the same, single “employer”.27 This approach to bargaining unit determination
26
In fact, policy responses aimed at this liability-avoidance consequence is the predominant concern in Weil’s
book. For criticism of this limited approach to assessing the consequences of fissuring, see Bartkiw (2015).
Similarly, see also the discussion of policy dialogue around the regulation of temporary agency employment and
how it has focused primarily on user firm avoidance of minimum employment standards and related liabilities, in
Bartkiw (2009), and fails to take into account shifts in worker bargaining power produced by triangular
employment relations (on this see also Bartkiw, 2014).
27
Of course, the “related employer” provision in section 1(4) of the Act “stretches” the default rule to some
degree. For a discussion of the resiliency of the default assumption in law that an “employer” be a single entity,
and how this is the historical residue of master and servant law, and its dependence upon a psychological
conception of a unitary “master”, see Fudge (2007).
31
Collective Bargaining, Strikes and Lockouts
magnifies the potential gains to lead firms from organizational fissuring strategies.
Fissuring produces further disaggregation in bargaining unit structure, pushing the
bargaining process to comparably lower “levels” throughout the economy (Bartkiw,
2015). Returning to the relevance of this context to strike dynamics, the
decentralization of bargaining structure re-maps the boundaries of permissive strike
activity. Fissuring generally narrows the scope of potentially disrupted work activities by
a strike of a given bargaining unit, and thus, under fissuring, lead firms (or firms “higher”
up the network) may be better able to adjust production activities in response to strike
disruption at a subordinate firm, by substituting an alternative supplier to the network.28
Some past quantitative studies may shed some light on the consequences of industrial
restructuring and fissuring on strike activity. In his econometric analysis of strike data
pertaining to Canadian manufacturing firms from 1965-1985, Budd (1994) found
evidence that smaller bargaining units had a lower propensity to strike than larger ones.
Using data on Ontario contract settlements from 1984-1993 in bargaining units of all
sizes, Hebdon at al. (1999) also found that small bargaining units had comparably lower
propensities to strike. These effects held in both the public and private sectors. They
also noted a trend towards smaller bargaining units in society over time, noting that the
average size of a sample of newly organized units was 48 employees, while their
aggregate sample bargaining unit size was 138 employees, suggesting that this would
likely generate decreases in strike activity in future. In terms of measures of the
probability of an impasse, their analysis suggested that the likelihood of a stoppage in
the smallest size categories was 8.8 percentage points lower than in the largest size
categories. Campolieti et al. (2005) also found that smaller bargaining units had a
lower probability of strike incidence. Notably, they found that the odds of a strike were
at least 82% smaller in bargaining units with fewer than 21 members than in those with
28
This substitutability effect likely explains why there does not appear to have been any successful attempts to
organize bargaining units of temporary help agency workers at the “level” of the individual temporary help agency
in Canada. See Bartkiw (2009).
32
Collective Bargaining, Strikes and Lockouts
500 or more members (Campolieti et al., 2005, at 627). These effects of unit size were
generally confirmed in Campolieti, et al. (2014).29
One specific form of workplace fissuring in recent decades has been the significant
expansion in temporary help agency employment (Vosko, 2000, 2010; Bartkiw, 2009).
This trend has a few potentially relevant implications on work stoppages activity, aside
from its effects on union growth in general.30 First, the expansion of agency-supply
arrangements as an increasingly normalized substitute for direct employment relations
has produced a larger, stable structure for indirect labour supply. The Canadian (and
Ontario) temporary help sector grew significantly in the 1990s and 2000s in terms of
nominal volumes of work, and in proportion to total labour supply (Bartkiw, 2009).
Although the precise question has not been examined, it suggests the possibility of an
increased capacity of employers to utilize agency-supplied labour in response to work
stoppages. Further, there is some evidence of an increasingly regular employer
practice in recent decades of building agency-supply arrangements into their normal
work operations, to varying degrees of intensity, with employers maintaining both a pool
of regular “direct hire” employees and a separate pool of agency-supplied workers
(Bartkiw, 2012). Assuming workers overcome the barriers to unionization that these
arrangements pose, the presence of agency-supplied workers performing similar or
related work to bargaining unit employees within an organization still threatens to
undermine strike efficacy, since the employer already has in place a pool of nonbargaining unit workers (unless the union has been able to achieve their full inclusion in
the bargaining unit) performing related/similar work as the bargaining unit, buttressing
29
Certain studies suggest some possible relationship between bargaining unit size and strike duration. Hebdon et
al. (1999) found a negative correlation between unit size and strike duration. Campolieti et al. (2005) found a
similar relationship. Finley (2010) “corrected” U.S. strike data by removing all strikes of a pre-determined length,
or strikes in which the precise length of it is established by the union before they occur. These authors do this
based on the theoretical assumption that in these strikes, the role of size does not bear upon duration, or does not
do so in a similar way as in cases where duration is not predetermined. Finley found that strikes tended to be
shorter when the striking workers unit represented a larger proportion of the employer’s overall workforce, and
that, with the inclusion of these new variables, the effect of raw bargaining unit size was no longer significant.
30
These additional effects on union growth may involve multiple causal pathways. See Vosko (2000); and Bartkiw
(2009, 2012, and 2014).
33
Collective Bargaining, Strikes and Lockouts
its ability to continue operations in the event of a strike (Bartkiw, 2012).31 Since these
arrangements are visible to bargaining unit employees in advance, they may reasonably
be expected to produce some strike deterrent effect.
Overall, there seems to be both theoretical and empirical support for the notion that
various contextual factors have produced a general decline in the instrumentality of
strikes, which helps explain the long run decline in strike incidence that we have
observed in recent decades.
B) The effects of specific components of the collective bargaining legal
regime on work stoppages and related outcomes
Another body of literature uses quantitative research methods to examine the
correlation between the existence of specific components of the legal regime governing
collective bargaining on various outcomes such as work stoppage incidence, duration,
total work days lost, wages, employment and investment. A few general comments on
this body of quantitative studies is needed, prior to reviewing the various results they
offer.
Most quantitative studies of strikes are self-described as being of a purely empirical
nature, in the sense that they are not framed as tests of any specific model of strike
behaviour per se. This is because many strike theories only provide ambiguous
predictions as to the effects of many policy variables, and there has not been significant
effort in incorporating policy variables into theories of strike behaviour.32 Most of these
econometric studies build upon the approach taken in a 1989 study analyzing work
stoppage data by Gunderson, Kervin and Reid. This study sought to examine the
independent effects of different components of labour policy using quantitative methods.
Intuitively, the models are designed to find policy-based explanations for the differences
in patterns of strike incidence (the dependant variable) across jurisdictions and time.
31
In their organizing activities, unions commonly need to make a strategic decision to exclude agency-hired
workers from proposed bargaining units in their certification application, despite commonality in job functions,
because of the comparable difficulty of obtaining support from amongst agency-hired workers, and the necessity
of majority support amongst workers in the bargaining unit established in the certification process. See Bartkiw
(2012).
32
See also the discussion above on the variety of theoretical models of strike behaviour.
34
Collective Bargaining, Strikes and Lockouts
The structure of this study, and all subsequent ones reviewed in this section, was “timeseries cross-sectional,” since data observations occurred in all Canadian jurisdictions
(10 provinces plus the federal jurisdiction), and also occurred over time (here, during the
period 1967-1986). All of the data in this study was limited to Canadian data, and the
same is true for all subsequent quantitative studies reviewed in this section. As part of
the quantitative modelling exercise, a set of “independent variables” is constructed to
represent, within the model, specific components of labor policy. At all points in time
and in each jurisdiction (Ie. whenever and wherever each contract may have been
settled), each of these labour policy independent variables are defined in a binary sense
as either existing or not (Ie. mathematically, to be equal to 1 or 0). A crucial precondition for whether this sort of analysis is even possible to perform (let alone
insightful), is that labour policy must vary. In fact, each and every separate component
of labour policy must vary, in order for it to even be possible to use any such component
of labour policy as a separate variable in the model. This is because all that the model
is ultimately doing with respect to each dependant and independent variable is trying to
measure whether the two variables are tending to move in some consistent manner
(correlation), or not. The main reason why there is a fairly large set of Canadian studies
of the relationship between strikes dynamics and labour policy (and so few U.S. ones) is
that researchers take advantage of labour policy differences across Canadian
provinces, treating this as variance. The amount of variation in labour policy over time
is not great.
In the original Gunderson Kervin and Reid (1989) study, they defined the following set
of 9 labour policy variables, and “coded” them based on their existence at precise timeperiods, and within each jurisdiction: conciliation (in either a 1-stage or 2-stage form);
cooling off periods (post-conciliation); mandatory strike vote requirements; employer
initiated (“final-offer”) votes; mandatory dues checkoff; anti-temporary replacement
worker restrictions; rights to reopen negotiations; rights to reopen negotiations due to
technological change. Subsequent studies of strike dynamics, although they are largely
built on the same time-series cross-sectional model structure, have altered this original
set of labour policy variables somewhat, which in itself can have an impact on some of
the estimates. This makes it hard to merely compare the results as representing
35
Collective Bargaining, Strikes and Lockouts
differences in policy effect shifting over time, and there are numerous conflicting results
for several labour policy variables across the different studies, which to some extent
then forces the reader to identify either technical or theoretical grounds on which to
prefer the results of one study over another.
In quantitative analysis of this sort, the measures of the separate coefficients on the
policy variables are estimates of the isolated independent correlation occurring between
the existence of the policy in question and the dependant variable being examined,
holding all other “effects” constant. In fact, although the word “effect” is sometimes
used in discussing results of quantitative studies, this is problematic terminology, since
quantitative models generally33 do not confirm causality, only correlation. Causality is
ultimately a matter of theoretical interpretation.
Since the particular historical period analyzed in a given study is also an important issue
to consider when assessing the evidentiary value and relevance of past studies, the
historical timeframes, jurisdictions, and data used in the various studies reviewed in this
subsection are summarized for quick reference and comparison in Appendix 2.
Although there are specific concerns pertaining to specific studies, which are discussed
below, a few concerns about the degree of reliability of this entire set of studies overall
should be noted at the outset. First, since a large part of the variation in labour policy is
actually cross-sectional differences, this means that capturing the effect of any such
differences effectively is important. However, there are limits to the ability to do this in
this approach. Models have limited ability to capture the various “province-specific”
characteristics or phenomenon that could also be affecting the various strike dynamics
in question. To the extent that those effects are not captured, models may produce
misleading estimates of the effects of labour policy variables, as the model “does not
know what it does not know”. Second, the components of labour policy do not exist in
isolation from each other, and are often adopted as large bundles of different provisions
simultaneously. This means that when a legislative change in a specific provision
represented by an independent variable is coded into a model, the model may take no
33
Some more advanced techniques are designed to improve causal inference, including instrumental variable
estimates, difference-in-difference estimates, regression discontinuity estimates, longitudinal estimate.
36
Collective Bargaining, Strikes and Lockouts
account of the fact that various other legal provisions were also changed at the same
time. Whether this “matters” or not depends on the specific context. When two
independent variables change at the same time, this reduces the ability of the model to
separate out the correlation between these variables and the independent variable.
This is an important issue in these studies, particularly in the case of trying to determine
the effects of anti-temporary replacement worker (“ATRR”) provisions. The specific
concerns for different models are outlined more fully in the discussion of the specific
outcomes of the studies below.
A third concern pertains to the essence or true state of
the law and the true differences in the law that are represented within the construction
of the labour policy variables in these models. In quantitative modelling, some
simplification of complex reality is necessary, and this is also the case in most forms of
social research. However, policy analysis must be cautious and mindful of the fact that
this simplification process is occurring. In this vein, there appear to be some conflicting
understandings of some aspects of Canadian labour law that have been “fed” into these
models by different researchers, and it is suggested that different researchers could
reasonably disagree on the appropriateness of including certain variables based on the
alleged degree of difference in the state of the law across Canada with respect to some
labour policy provisions. More detailed examples of this issue are provided in the
review of specific studies below. Overall, these concerns about model specification
raise significant concerns about the ability to rely on the results of these models to
justify policy choices relating to specific components of labour policy.34
In what follows, a fairly detailed review of these studies is provided. Rather than
proceeding through each study, one at a time, and reviewing all of the results pertaining
to each policy variable in each study, the results of these studies have been organized
into separate subsections for each of the labour policy variable sub-categories
requested by the CWR in the Statement of Work. Specific concerns relating to specific
studies are also discussed below.
34
Some scholars have expressed skepticism about the ability of econometric models to incorporate the necessary
complexity in real-world factors determining work stoppages and related outcomes. See for example, Savage and
Butovsky (2009).
37
Collective Bargaining, Strikes and Lockouts
i)
Anti-temporary replacement worker restrictions (ATRR)
In the initial study of 1967-86 unit-level contract settlement data, Gunderson, Kervin,
and Reid (1989) found that the relationship between ATRR and strike incidence was
statistically significant, and their model suggested that the presence of ATRR in a
jurisdiction was associated with a 24.4 percentage point increase in the probability of a
strike in a given round of negotiation in that jurisdiction.35
Gunderson and Melino (1990) also examine the correlation between various labour
policy variables, including ATRR, and strike duration, using data from 1967-85, but with
data restricted to large (500+ employee) bargaining units. They found that ATRR was
significantly associated with substantially longer strikes. However, the authors noted
that reliance upon this result requires caution since it was not only surprising, but was
also based entirely on the experience of Quebec, and thus the model might be picking
up other Quebec-specific effects.
Budd (1996) found that in one of his models, the effect of ATRR on strike incidence was
statistically significant and positive, increasing the likelihood of a strike by 13.3
percentage points. However, pointing out that there may be other province-specific
effects in the data that are not fully captured by the model, Budd also ran separate
models on the data with provincial dummy variables attempting to control for these
differences. The inclusion of these controls greatly reduced the size of the estimated
effect of ATRR on strike incidence, and although they remained positive, the estimates
were no longer statistically significant, meaning that they were less precisely estimated,
and there was insufficient evidence to reject the null hypothesis of their having no effect.
Budd also tested for the effect of ATRR on strike duration. He found that ATRR was
significant and associated with longer strikes. Lastly, Budd examined the relationship
between ATRR and wages, using various indicators of wage growth. Here, he found
conflicting results (positive and negative outcomes) using two different wage measures.
Further, once province-specific effects were controlled for with dummy variables, the
correlation with ATRR became much smaller in size and ceased to be statistically
35
This value is from their Table 2, which provides three separate estimates based on underlying strike probability.
I use the results in their column 5, estimates based on underlying average strike probability.
38
Collective Bargaining, Strikes and Lockouts
significant. Budd’s study illustrates the difficulty of analyzing the relationship between
ATRR and other variables, since there is a limited set of historical observations
available, with studies up to 1993 relying almost entirely upon observations pertaining
only to Quebec, with limited ability to analytically isolate the independent effect of this
one policy from all other province-specific phenomena.
Langille (1995) discusses Ontario’s ATRR provisions (in effect during 1993-95) and
comments on recent studies (Budd, 1996) of the effect of this short-lived policy change.
He questions whether Ontario’s 1993-95 ATRR provision can be presumed to have had
much of an effect, in any dimension. He argues that quantitative analysis seems to
merely assume it is possible and appropriate to model the effect of this legal change
with insufficient theoretical attention as to whether the new law intuitively threatens to
alter actor behaviour to a significant degree. In minimizing the significance of the 1993
ATRR provision, he cites the low volume of historical usage of replacement workers
(approximately 12% in Ontario in the year 1991 according to Haywood (1991)) and
suggests that this supports the claim that the ATRR provision likely made little practical
difference, perhaps contributing to the difficulties of capturing statistical significance of
the ATRR variable in quantitative studies.36
Crampton, Gunderson, and Tracy (1999) examine the correlation between ATRR and
strike incidence, duration and wages from 1967 to 1993. Their results were similar to
prior studies, but yielded more modest measures of the size of relationships. Their
models estimated that the presence of ATRR increased the probability of a strike in a
given case by 12.2 percentage points, although the effect was not found to be
statistically significant. Further, ATRR was significantly associated with much longer
strikes, with sizes varying by choice of model specification. In their preferred model,
ATRR was associated with a 46% increase in average strike duration. As well, in their
models of real wages, ATRR was associated with slightly more than a 4% increase in
real wages.
36
However, this does not address the question of the threat of using replacement workers, under alternative rules.
39
Collective Bargaining, Strikes and Lockouts
Budd (2000) examined the correlation between different labour policy variables,
including ATRR, on provincial employment, using data on monthly observations from
January 1966 to December 1994. This period of study allowed for observations not only
during the timeframe when ATRR was in place in Quebec, but also for when it was in
place briefly in Ontario and BC, although it only includes a portion of the period during
which ATRR was in place in Ontario and BC. A fairly simplistic “difference-in-difference”
exercise, which used growth in other provinces as a benchmark but did not attempt to
control for other potentially relevant causal variables, suggested that ATRR was
associated with a decrease in the aggregate employment-to-population ratio in each of
Quebec, Ontario and B.C. Regression models were also constructed, analysing trends
in the employment-to-population ratio and in unionized employment, and all suggested
that ATRR was significantly associated with a reduction in employment. Different
models suggested that ATRR was associated with a decline in the employment-topopulation ratio (which had a mean of 57.9 percent) by .47 percentage points to 1.28
percentage points. Using a different measure of employment, other models also
suggested that ATRR was significantly associated with fairly substantial declines in
unionized employment. Models that attempted to control for province-specific effects to
some degree resulted in smaller estimates of employment decline, and there was
evidence that time-varying effects were important. Unfortunately, Budd’s overall
modelling design did not allow for estimating the effect of province and year effect
interactions – a technique subsequently performed on various policy variables (including
ATRR) in Campolieti et al. (2014), and with important consequences for the results.
Budd and Wang (2004) analyzed the correlation between striker reinstatement rights
and ATRR on provincial investment rates. Their data included observations of
provincial net investment rates and related variables from 1967 to 1999. They found
that net investment was significantly and negatively correlated with ATRR. In terms of
size, their models estimated that ATRR was associated with approximately a 25%
decline in net investment. In models assessing the degree of correlation between
ATRR and investment across different industries, they found that the comparably
strongest relationship appeared to be in building construction. In models of the timevarying effects of labour policy, only the negative effect on building construction (and no
40
Collective Bargaining, Strikes and Lockouts
other industrial sectors) was significant in the first three years after ATRR took effect,
but an overall negative effect was also estimated for the period following three years
after the adoption of ATRR.
Duffy and Johnson (2009) evaluated the effect of ATRR in a more recent dataset from
1978-2003, and used annual, province-level aggregated data rather than unit-level data
used in most other studies. The design of this study on the one hand used somewhat
less information, given that data was more aggregated, but on the other hand involves
somewhat more historical variation in the use of ATRR, reducing somewhat the
dependence on the historically specific outcomes observed only in Quebec. Further,
following Budd (1996) they employed additional labour policy variables attempting to
further control for the effect of the presence of striker reinstatement rights (and thus
bans on permanent replacement workers), and of the presence of strike-breaker
restrictions. They found that the presence of ATRR was significantly correlated with an
increase in strike incidence by approximately 50%. Contrary to prior studies, they also
found the presence of ATRR was significantly associated with a large decline in strike
duration. With the mean strike duration of 34.14 days, two models yielded estimates of
reductions of duration by 35 and 49 days. Lastly, they found that ATRR was
significantly associated with an overall increase in total days lost to work stoppages,
with two models yielding estimates of 880 and 1550 days lost per year. Duffy and
Johnson also test to see whether the correlation between ATRR and these three
different phenomena (incidence, duration and total days lost) vary over time, and
specifically look for a difference in effect between the period of 1-2 years before ATR
was in effect, the 2 year period after the legislation is passed, and the period more than
2 years after the legislation took effect. Their results suggest that with respect to
incidence and duration, the effect of ATRR persisted for more than the 2 year period,
but that there was very weak evidence that this translated into an increase in days lost
to work stoppages, after the 2 year period.
Dachis and Hebdon (2010) use unit-level data from 1978-2008, and a similar set of
policy variables as used in prior studies in this area. They found a significant positive
effect of ATRR on strike incidence; with a mean of .72 strikes per month per province,
41
Collective Bargaining, Strikes and Lockouts
their models suggested that ATRR increased strike incidence by .11 strikes per month,
roughly a 15% increase. They further found that ATRR was associated with a 58%
increase in average duration, and that the effect on duration was significant. Dachis
and Hebdon also examined the effect of ATRR on wages. Contrary to their
expectations and results of some prior studies, they find that ATRR had a significant
negative effect on wages, with their models yielding an estimated reduction of 3.4% in
the level of private sector real wages. A concern about the results in this study is that
most of the models in the study of the behaviour of numerous dependant variables,
were run using a standardized set of labour policy variables, such that models of strike
dynamics also included an additional independent variable representing whether a
secret ballot union certification process was required in existence in each jurisdiction
and time period, with arguably questionable theoretical justification.37 Such a variable
would be more properly included in studies of union growth, and seems out of place in
these models. This specification choice may also affect reliability of other results in this
study to some degree.
Campolieti, Hebdon and Dachis (2014) perform quantitative analysis on strike data from
1978-2008, using unit-level data pertaining to bargaining units of all sizes. This is the
most recent econometric study of the effect of ATRR, using the most recent dataset.
Also, the authors perform certain improvements in estimation technique, not included in
prior papers. One key improvement involves correcting for “clustered standard errors”,
which are the result of using labour policy variables that vary only at the province level,
in an analysis of unit-level data. Employing this correction caused fairly significant
changes to results from prior studies. Many estimated coefficients were found to be
smaller in size, and statistically insignificant. A key difference in the results was that
they could no longer reject the null hypothesis that all labour policy variables tested as a
whole, had zero effect on strike incidence. Further, since the authors noted the
existence of major contextual shifts in the external environment in more recent decades
(discussed earlier in this report), they also examined whether there was evidence of
time-varying effects of different labour policy variables, on the various outcomes. To do
37
I would note that in Campolieti at al, 2014, a subsequent published study involving these same two authors as
co-authors, this variable was removed.
42
Collective Bargaining, Strikes and Lockouts
this, they used “dummy” variables for the time period after 1992, and looked to see
whether there was significant effect of various “interaction terms” measuring the
combined effects of these dummy variables and labour policy variables. In other words,
they looked to see if there was evidence of a statistically significant different effect of
policy variables like ATRR, on different outcomes, after 1992. With respect to strike
incidence, they found that ATRR also had no statistically significant effect in the post1992 period. They also estimated that ATRR was significantly and positively correlated
with strike duration, and that there also appeared to be a greater degree of correlation
between ATRR and strike duration in the post-1992 era. Certain tests they performed
also provided evidence that these effects were robust across bargaining unit sizes.
Lastly, their models of wage dynamics for the entire 1978-2008 time period suggested
that ATRR has been associated with a decline in wages of about 1.8% throughout this
period. However, their models also suggested that the correlation between ATRR on
wages has been highly time-sensitive. Using various different specifications, they found
that the correlation between ATRR and wages was statistically significant and positive
in the more recent, post-1992 era. Estimates of the size of wage growth associated
with ATRR post-1992 varied across models from 2.3% to 7.3%. The authors suggested
that one possible interpretation of these reversed effects of ATRR (and other labour
policy variables) is that in more recent years, policy functionality has changed as a
result of the new environment of expanded employer bargaining power.
Lastly, Legree, Schirle and Skuterud (2014) estimate the effect of various components
of labour law on provincial union density rates over time. In their estimates of the
isolated independent effects of various components of law, they found that ATRR was
significant and associated with a positive effect on union growth, and in fact had a larger
estimated effect than all of the other components of law that they examined, including
the type of union certification procedure.
Comments and Assessment of these Studies:
In summary, with respect to the crucial policy question of what have been the historical
effects of ATRR provisions in Canada, we have somewhat of a range of answers from
the literature published over the past 36 years.
43
Collective Bargaining, Strikes and Lockouts
With respect to strike incidence, prior to Campolieti et al. (2014), there was near
consensus that ATRR increased strike incidence, but with estimation improvements in
the latter study, this older consensus now seems rebuked. With respect to strike
duration, most studies suggested that ATRR was also associated with increased strike
duration, although Duffy and Johnson (2009) found the opposite. Under estimation
improvements in Campolieti et al (2014), ATRR was found to be positively correlated
with strike duration to a comparably greater degree in the more recent (post 1993) era.
With respect to wages, some early studies suggested that ATRR was associated with
higher wages, suggesting that this operated through a shift in bargaining power,
although Budd (1996) and Dachis and Hebdon (2010) produced conflicting results,
suggesting a negative wage effect operating through investment reductions. Budd
(2000) and Budd and Wang (2004) support this latter interpretation since their results
suggested provincial aggregate investment was negatively correlated with ATRR.
However, with their improved estimation techniques and most recent data, Campolieti et
al. (2014) find that ATRR is associated with wage increases in the most recent (post1993 era).
Overall, there appear to be various methodological grounds for preferring the results of
Campolieti et al., 2014 over other studies, although there are still certain concerns about
this study, along with the rest, that are reviewed below.
As admitted in most of these studies, labour policy provisions are typically not adopted
in isolation, but rather as part of a legislative package, with a bundle of components.
With respect to ATRR, even the most recent econometric studies reviewed above are
heavily dependent, in a mathematical sense, upon measures of outcomes observed
during a short-lived presence of ATRR in Ontario from 1993-1995. However, the ATRR
mechanism that was in place in that short period was part of a large package (“Bill 40”)
of labour law reforms, many of which also had a fairly short lifespan, since they were
also revoked with the adoption of “Bill 7” in 1995.38 Assuming then that it is even
possible to speak of purely independent effects of ATRR alone, separate from all other
aspects of law, these models are not able to isolate the “effects” caused by other
38
Bartkiw (2008) includes a summary of the different shifting components of labour law in this timeframe.
44
Collective Bargaining, Strikes and Lockouts
components of labour law that also shifted in Ontario during this timeframe, from the
effects of ATRR, because they shifted together as a bundle. This means that, assuming
there were investment and employment effects experienced in Ontario that were truly
due to Ontario labour law shifts in the mid-1990s, then the models ought not to be
interpreted as picking up evidence of effects of ATRR alone.
Relatedly, as was discussed earlier in this paper, there is a general limitation across
these studies that labour policy variables themselves do not actually vary (the essence
of a variable) all that much. Much variation is cross-sectional, while variation over time
is minimal, compared to the behaviour of many other commonly used variables, like
GDP growth or other business cycle indicators. This affects the ability to “zero in” on
the true relationship between a specific labour policy and other phenomena, and means
that results produced will be increasingly affected by how models are specified.39
Returning to our uncertainty over the question of employment and investment related
effects of ATRR, if we look at the research design in Budd (2000) and Budd and Wang
(2004), these models, out of necessity, rely on the Ontario experience of labour policy
change in the mid-1990s to a substantial degree. That is, the estimated relationships,
and their significance, are based in no small part on the fact that in the mid-1990s, while
ATRR was in place in Ontario, there were observed declines in investment and
employment levels. However, during much of this same time period, the Ontario
economy was also being affected by a fairly widespread recession, and both investment
and employment are expected to decline during a recession. Matters are further
complicated by the fact that the North American Free Trade Agreement took effect on
January 1, 1994, and this may have had certain additional investment and employment
related effects due to industrial restructuring in the Ontario economy. To try to separate
out both of these business cycle and industrial structural effects from labour policy
effects, Budd (2000) and Budd and Wang (2004) do use various control variables
39
To a degree, this difficulty was seemingly masked in some of the prior studies in this area that did not employ
the technique of correcting for “clustered standard errors” that are generated with the combined use of unit-level
data on certain variables with provincial-level data on labour policy variables, the latter of which suffer from a lack
of variation. This technique which effectively “punishes” outcomes on this basis, making estimated results more
conservative, was only applied in Campolieti et al. (2014). Apparently, this technique was not technically feasible
at the time of most of the earlier studies, but has become feasible and reasonably expected to be done in such
studies, in recent years.
45
Collective Bargaining, Strikes and Lockouts
representing different measures of the business cycle, and they also include controls for
the proportion of employment represented by manufacturing and public administration.
Ultimately however, the fact is that these employment and investment declines took
place at a time in Ontario when declines of some size were to be expected from the
recession (and possibly also from industrial restructuring). Given a general lack of
variation in labour policy over time for comparison to investment/employment
fluctuations over time, this raises concerns about the reliability of model estimates of the
specific proportion of variation in employment/investment that is correctly attributed to
labour policy as opposed to other factors. In this modelling context, the degree to which
employment/investment variation is attributed to labour policy as opposed to other
factors will be sensitive to the precise econometric model specifications chosen by the
researcher.40
ii)
The legislative requirements for a legal work stoppage:
Many of the quantitative studies already reviewed in the above section also sought to
estimate the effects of three different kinds of legislative requirements for a legal work
stoppage: mandatory conciliation, mandatory strike votes, and mandatory cooling off
periods.
Gunderson Kervin and Reid (1989) found that mandatory conciliation was associated
with reduced strike incidence, with a larger and more significant reduction found in
relation to two-stage conciliation. Their models also suggested that an additional
cooling-off period after conciliation had little association with strike incidence, and that
mandatory strike votes had a significant and large effect on reduced strike probability.
Gunderson and Melino (1990) found that the relationship between conciliation (in either
one-stage or two-stage forms) and strike duration was not significant. The cooling off
period variable was significant and was associated with a reduction in duration. Also,
the presence of mandatory strike votes was significant and was associated with
substantial reductions in strike duration.
40
This includes choices about the set of variables to include; different data measurements or indicators pertaining
to certain variables; whether certain variables should be linear, logarithmic, or lagged; and the estimation
procedures chosen.
46
Collective Bargaining, Strikes and Lockouts
Crampton, Gunderson and Tracy (1999) estimated the effects of each of these three
policy variables on strike incidence, duration and wages. With respect to strike
incidence, they found evidence that conciliation reduces strike incidence, and that twostage conciliation produces a larger effect, although their estimates were not statistically
significant. Under their estimates, a 14 day cooling off period reduces strike incidence
by 4.8 percentage points compared to a jurisdiction without conciliation. Similarly,
mandatory strike votes were significant, and associated with a reduction of over 7
percentage points in strike incidence. In their models relating to strike duration, they
found that conciliation was associated with longer strikes, suggesting that its success in
reducing strike incidence may pertain to what would have been shorter strikes, shifting
the composition of strikes that do occur. Both mandatory strike votes and cooling off
periods were associated with shorter strikes, but only the former was statistically
significant. With respect to wages, only the cooling off period variable was found to be
significant, and their models suggested that it was associated with quite a large
reduction (4%) in wage settlements. This seems to be a rather radical finding. Since
there is no apparent theoretical justification for why this particular policy instrument
should influence wage levels to such a great extent (in either direction) this outcome
therefore reasonably raises concerns about accuracy in overall model specificity.
Budd (2000) includes control variables for mandatory strike votes and cooling off
periods in his models of the effects of ATRR on employment, although he does not
include conciliation as a variable. He finds that mandatory strike votes are significantly
associated with a small increase in the employment rate. With respect to cooling off
periods, the evidence is mixed, with them being significant and negatively correlated in
some models, but insignificant in others. Again, there is debatable theoretical
justification for inclusion of these bargaining process variables, and a theoretical case
for their causal effect on employment levels is not even provided by the author. Rather,
Budd simply notes that sometimes his main variables of interest (replacement rights and
ATRR) are adopted as part of a package of other measures, and so the inclusion of
these other bargaining process variables in this model of employment is done to try to
isolate out their effects, but again, this begs the question of whether any effects are a
47
Collective Bargaining, Strikes and Lockouts
priori expected. In this situation, relying on the estimates pertaining to these control
variables as accurate measures of employment effects demands great caution.
In their models of strike incidence, Duffy and Johnson (2009) found that neither
compulsory conciliation nor mandatory strike votes were significant. The number of
days of a cooling off period, however, was significant, and incidence declined with
increases in the cooling off period. In their models of work stoppage duration, they
found that conciliation was not significant, but some evidence that cooling off periods
were significant and reduced strike duration, while mandatory strike votes were
significant and associated with increased average strike duration of 28 days. This latter
result seems particularly surprising and conflicts with results in prior studies. In their
models of days lost to work stoppages, neither cooling-off periods nor mandatory strike
votes are significant, but in one model, conciliation is significant and associated with
very large reductions in days lost.
In their models of wage effects, Dachis and Hebdon (2010) find that conciliation was not
significant, but mandatory strike votes and cooling-off periods are significant and
associated with substantially higher (2.8%, and .4% per day respectively) real wage
levels. These are somewhat unexpected results, there being little a priori theoretical
justification for such substantial wage effects of these particular bargaining process
variables. In their models of strike incidence, conciliation is significant and associated
with fewer strikes, while mandatory strike votes are also significant but associated with
more strikes. The authors omit any reference to the relationship between cooling-off
periods and strike incidence. In their models of strike duration, they find that mandatory
conciliation and cooling-off periods are significant and associated with substantial (41.2
days and 2.5 days/day of cooling-off respectively) increases in average strike duration.
They also found that mandatory strike votes were significant and substantially reduced
strike duration. All of these results concerning strike duration conflict with results in
Duffy and Johnson (2009). Lastly, see the discussion above about it arguably being a
model specification error in this study to include a variable relating to the presence of a
secret-ballot vote in the certification process as a variable in these models, affecting the
reliability of results in this study.
48
Collective Bargaining, Strikes and Lockouts
In contrast to most of the prior research in this area, Campolieti et al (2014) found that
each of conciliation, mandatory strike votes, and cooling-off periods are not significantly
correlated with strike incidence. Their models controlling for time-varying effects
suggest that the effects of cooling off periods in incidence may be larger pre-1993, but
the effects of conciliation and mandatory strike votes are larger post-1993. In their
models of strike duration, conciliation is not significant, while cooling off periods were
significant and associated with a small increase in strike duration. Mandatory strike
votes are also significant and associated with shorter strikes. Their models testing for
time-varying effects, suggest that the effects of mandatory strike votes were larger in
the pre-1993 data, the effects of conciliation are larger in the post-1993 data, while the
effects of conciliation were of roughly the same size, but in the opposite direction. In
their OLS models of wage effects, the presence of conciliation was associated with a
1% decline in real wages, cooling off periods were associated with a 0.2%/day increase
in wage settlements, while mandatory strike votes were not significant. In models
testing for time-varying effects, the effect of mandatory strike votes was larger in the
pre-1993 data, while the effects of conciliation are larger in the post-1993 data.
Comments and Assessment of these Studies:
Studies in this area seem to have produced some conflicting results, and indeed some
rather surprising results with respect to the effects of these three labour policy variables.
A significant concern is that there appears to be little a priori theoretical grounds for the
implicit assumption made in many of these studies that in models of wage
determination, these three dispute-narrowing process variables should also be included
as independent variables. Nevertheless, it is a fairly clear pattern that most of these
studies simply opted to include the entire set of labour policy variables that were initially
used to model strike incidence in Gunderson et al (1989). While these may have been
initially appropriate in theoretical terms for modelling strike incidence and/or duration,
the determination of wages, employment, and investment are phenomena that are
comparably more distant from these dispute-narrowing process variables, which
reduces the theoretical justification for their inclusion. Their inclusion seemingly also
resulted in some important effects on the results, since large wage effects seem to have
49
Collective Bargaining, Strikes and Lockouts
been attributed to one or more of these policy variables in certain studies, which on
theoretical grounds is difficult to accept as being reliable.
With respect to the effects of these variables on strike incidence and duration, as noted
above, on methodological grounds, there are a number of reasons to prefer the results
obtained in Campolieti, et al. (2014) over others, and their findings suggest comparably
more modest effects of these three labour policy variables. Somewhat surprising
findings even in their study included a positive correlation between cooling off periods
and strike duration. Given the presence in the data of what appears to have been a
positive trend in strike duration in recent years, it is possible to interpret this result, and
possibly some of the other findings pertaining to duration, as suggesting that the
presence of these policy variables is not causing longer duration, but that policy is
increasingly incapable of producing settlement in cases involving larger zones of
disagreement (and/or that there are proportionately more of such cases) leading to
longer strikes being observed, while still being able to settle “smaller” conflicts that
would have resulted in shorter strikes.
iii)
Reinstatement of employees following a work stoppage
A number of studies already discussed in this review also include analysis of the effects
of striker reinstatement rights, such as the provision in section 80 of the Labour
Relations Act, 1995. These rights effectively create a ban on the use of so-called
“permanent replacements” by employers.41
Budd (1996) finds that reinstatement rights were not significantly correlated with strike
incidence, and nor were they significantly correlated with strike duration. In his models
relating to wage effects, reinstatement rights were generally statistically significant, with
certain models suggesting a negative but negligible effect on wages, while other models
suggested a larger negative effect of around -1%.
Budd (2000) models the effects of reinstatement rights, inter alia, on employment, with
conflicting results. Certain of his models suggest that reinstatement rights are
41
Employers in the U.S. are generally entitled to use permanent replacements, with certain provisos. See NLRB v.
Mckay Radio and Telegraph Co., 304 U.S. 333, 346 (1938).
50
Collective Bargaining, Strikes and Lockouts
significant and associated with an increase of roughly .33 to .55 percentage points in
the provincial employment-to-population ratio. Other models suggest that reinstatement
rights are significantly associated with small declines in unionized employment, with
insignificant effects over time.
Budd and Wang (2004) model the effects of reinstatement rights, inter alia, on provincial
investment. Their models suggest that the presence of reinstatement rights was not
sufficiently associated with aggregate net investment, although some models suggested
a significant and negative relationship between these rights and investment in the
building construction subsector.
Duffy and Johnson (2009) found that reinstatement rights were significant and
associated with an increase in strike incidence. In their models of strike duration,
reinstatement rights were also significant and associated with a very large decline (45
days) in average duration. Both of the effects on incidence and duration were found to
persist for more than 2 years after legislation was enacted. In their models of days lost
to work stoppages, reinstatement rights were not significant.
Dachis and Hebdon (2010) found that reinstatement rights were significant and
associated with very large declines (-5.3%) in private-sector real wages; were
significantly associated with higher strike incidence (.44 extra strikes per month per
province) and much longer strikes (an extra 47.4 days on average). See the discussion
above about it arguably being a model specification error in this study to include a
variable relating to the presence of a secret-ballot vote in the certification process as a
variable in these models, affecting the reliability of results in this study.
Campolieti et al (2014) found a small positive relationship between reinstatement rights
and strike incidence, the effect was not significant, and the size was larger post-1992.
In their models of strike duration, they found that reinstatement rights were significant
and associated with substantially longer strikes, with conflicting estimates of timevarying effects. In their models of reinstatement rights, they found that these rights
were significant and associated with estimated declines in wages of 1.7% to 3.8%, with
larger effects post-1993.
51
Collective Bargaining, Strikes and Lockouts
Legree, Shirle, and Skuterud (2014) take the approach of treating the existence of
reinstatement rights and bans on permanent replacement rights as being separate
variables from each other. Previous studies treated all such rules as a single variable
contruct. This is the first study in this line of studies to treat these this, since qualitative
differences in the essence of these legal provisions is highly debatable. This raises
concerns about the reliability of their results.
Comments and Assessment of these Studies:
These studies produced some conflicting results with respect to these components of
labour policy. As noted above, on general methodological grounds, there are a few
reasons to prefer the results in Campolieti, et al. (2014) over earlier studies.
As noted above, there are additional concerns about these models. Beginning with
Budd (1996), these econometric studies began adding the presence of reinstatement
rights as an additional independent variable in models, on the notion that not every
Canadian jurisdiction provided these rights. Prior studies of labour policy did not include
the presence of these rights as separate control variables seemingly because the
authors either held a different interpretation of Canadian law, or because it was not
clear that the law in different provinces on this issue was actually sufficiently different in
a qualitative sense to justify treating the presence of these rights as being something
that really varied throughout Canada.42 Generally, provincial statutes tend to specify
that workers either have a right to reinstatement, or they alternatively contain some
protection to the effect that workers cannot lose their employment as a result of the
strike. Although formal distinctions may make it technically possible in extraordinary
cases for an employer to use replacements indefinitely, the law has a degree of
similarity, and other instruments, like unfair labour practice rules, may further narrow the
difference in practice. Nevertheless, following Budd (1996) most of the authors in this
area began to insert reinstatement rights as independent causal variables in their
models. Indeed, one recent study (Legree et al, 2014) in addition to a variable for
ATRR, there are separate variables used for each of reinstatement rights and bans on
42
In addition to studies noted already, see also Martinello and Meng, 1993.
52
Collective Bargaining, Strikes and Lockouts
permanent replacements, despite the near identity between the two concepts. In this
study, variables are also coded on the assumption that Ontario law has never contained
any ban on permanent replacements, even though Ontario has had reinstatement rights
since 1976, including its most recent form in section 80 of the LRA since 1995. If this is
not somehow a model specification error in this study, it at least raises serious concerns
about how results are to be interpreted.
Overall, if these concerns about these models based on contestable legal
understandings are accepted as valid, this raises caution not only about relying upon
the estimates of the various effects being attributed to the existence of reinstatement
rights in these models, but may also raise additional caution about relying on other
results produced by these models.
iv)
Mandatory and permissive provisions in collective agreements
A number of empirical studies already discussed in this section also attempt to estimate
the effects of legislative provisions that fall under this category: mandatory dues
deduction and reopener rights. The latter are sometimes treated as two separate
categories, since sometimes the law allows for general reopener provisions to be
included in collective agreements, while in some other situations the law specifically
provides some protection for reopener rights in the case of technological change.
Gunderson, Kervin, and Reid (1989) found that neither the allowance of general
negotiated reopeners, nor automatic technological reopeners were significantly
associated with strike incidence. They also found that mandatory dues deductions were
associated with a large and statistically significant reduction in strike incidence.
In their models of strike duration, Gunderson and Melino (1990) found that while general
reopeners were not significant, automatic technological reopeners were significant and
associated with a reduction of about 3.4 days in average strike duration. They also
found that mandatory dues deductions were significantly associated with an increase in
strike duration of about 6.4 days.
53
Collective Bargaining, Strikes and Lockouts
Crampton, Gunderson and Tracy (1999) found that general reopener rights were
significant and associated with fairly large reductions in strike incidence. On the other
hand, technological reopeners were significantly associated with fairly large increases in
strike incidence. In their models of duration, only general reopener rights were
significant, in the model pertaining to strikes in all bargaining unit sizes. In their wages
models, these provisions were not significant. With respect to mandatory dues checkoff provisions, these were found to be not significantly correlated with strike incidence,
but were significant and associated with large (50%) increases in strike duration. In
their models of wage effects, mandatory dues check-off provisions were not significant.
In his models testing for the relationship between replacement worker restrictions and
employment, Budd (2000) includes the presence of mandatory dues check-off
provisions as a control variable. In each of his models, the presence of mandatory dues
check-off provisions was insignificant.
Duffy and Johnson (2009) find that mandatory dues checkoff rules are significantly
associated with a substantial decline (of about 50%) in strike incidence. In their models
of work stoppage duration, mandatory dues checkoff provisions were not significant. In
their models of days lost to work stoppages, mandatory dues provisions were found to
be significant and associated with very large reductions in days lost. Duffy and Johnson
only include the control variable for technological reopener rights, and find that these
are not significantly correlated with work stoppage incidence or duration but that they
are significantly associated with a large increase in total days lost to work stoppages.
Dachis and Hebdon (2010) found that the presence of mandatory dues checkoff
provisions was not significantly correlated with wages, strike incidence, or strike
duration. Also, as noted above, it is arguably a model specification error that this study
included a variable relating to the presence of a secret-ballot vote in the certification
process as a variable in these models, affecting the reliability of results in this study.
Campolieti et al (2014) found that the presence of mandatory dues checkoff provisions
was not significantly correlated with strike incidence, strike duration, or wage levels.
They also found that the neither the presence of general reopener rights or
54
Collective Bargaining, Strikes and Lockouts
technological reopener rights was significantly correlated with strike incidence. In their
models of strike duration, they found that general reopeners were not significant, but
that technological reopeners were significant and associated with substantial increases
in strike duration.
Legree, Schirle and Skuterud (2014) found that neither the presence of mandatory dues
checkoff provisions, nor the presence of technological reopener rights, were significantly
correlated with union growth.
Comments and Assessment of these Studies:
These studies produced some conflicting results with respect to these components of
labour policy. As noted above, on methodological grounds, there are a few reasons to
prefer the results in Campolieti, et al. (2014) over earlier studies.
An additional concern about these studies is that different studies have seemingly used
different coding practices for the labour policy variables, based apparently on conflicting
understandings of the actual state of the law. For example, in Gunderson et al (1989),
Gunderson and Melino (1990), Crampton et al. (1999), and Duffy and Johnson (2009),
models are coded on the understanding that Ontario law has never provided reopener
rights, during the entire period of each study. In contrast, in Campolieti et al. (2014),
models are each structured on the contrary understanding that Ontario law has always
provided for general reopener protection throughout the entire historical period, but not
specific protection for technological reopeners. These conflicting legal interpretations
(and resulting model specification) not only make comparisons across the different
studies very difficult, but may also raise some concern about the ability to rely upon the
estimated effects of these specific legal provisions produced by these models, and/or
possibly raise some concerns about relying on other estimates produced by these
studies due to model specificity concerns.
55
Collective Bargaining, Strikes and Lockouts
v)
Other bargaining process policy measures – employer initiated “final offer”
votes
Some of the econometric studies already reviewed in this report also attempt to
estimate effects of the presence of employer rights to request “final offer” votes.
Gunderson Kervin and Reid (1989) found that these provisions were significantly
associated with substantially higher rates of strike incidence.
Gunderson and Melino (1990) found that these provisions were not statistically
correlated with strike duration.
Crampton, Gunderson and Tracy (1999) found that these provisions were not
significantly associated with strike incidence, but that they were significantly associated
with a large decline in strike duration. They also found that they were significantly
associated with a small increase in wages.
Duffy and Johnson (2009) found that employer initiated votes were not significantly
correlated with work stoppage incidence, nor total days lost to stoppages, but that they
were significantly associated with longer average work stoppage duration. See the
discussion above about it arguably being a model specification error in this study to
include a variable relating to the presence of a secret-ballot vote in the certification
process as a variable in these models, affecting the reliability of results in this study.
Dachis and Hebdon (2010) found that employer initiated votes were not significantly
associated with wages, nor strike incidence, nor strike duration.
Campolieti et al (2014) found that employer-initiated votes were not significantly
associated with strike incidence, nor duration, nor wages.
Legree, Schirle and Skuterud (2014) found that employer-initiated votes were not
significantly associated with union density rates.
56
Collective Bargaining, Strikes and Lockouts
Comments and Assessment of these Studies:
These studies produced some conflicting results with respect to the effect of employer
“final offer” votes. As noted above, on methodological grounds, there are some reasons
to prefer the results in Campolieti, et al. (2014) over earlier studies. Overall, the
evidence from these studies is that these provisions appear to be neutral in effect
towards the variables of interest.
Overall, as was discussed at the beginning of this section reviewing quantitative studies
of strike dynamics, and repeated throughout the summary of various studies, there are
a number of concerns about these studies collectively that demand significant caution in
relying upon their results as justification for policy change pertaining to specific
components of labour policy.
C) The Regulation of Picketing Activities
The regulation of picketing activities has deep historical roots in the common law, and
formed an important component of early regimes of labour regulation (Fudge and
Tucker, 2001). A significant amount of academic research is focused on tracing out the
historical roots and evolution of the law governing picketing. Much of this work seems
rather critical of the historical basis of picketing regulation, based as it is on a body of
nominate and economic torts, and certain provisions in the Criminal Code, developed in
the late 19th and early 20th centuries (Eaton, 1992; Adell, 2003; Smith, 2014).
Much academic analysis of picketing regulation highlights the role of judicial ideology in
the development of legal doctrine over time (Sangster, 2004; Tucker, 2010). Historical
analysis illustrates how on the one hand, judicial ideology is to some extent a product of
the larger political-economic context (Tucker, 2010). However, ideology has also
played a more specific role in precise historical contexts and moments. Tucker (2010)
illustrates an important example of such a juncture in his analysis of the historical
background to the 1963 decision of the Ontario Court of Appeal in Hersees of
Woodstock Ltd. v. Goldstein,43 in which the court held that secondary picketing was per
43
(1963), 38 D.L.R. (2d) 449 (Ont. C.A.).
57
Collective Bargaining, Strikes and Lockouts
se illegal. This decision involved some now infamous reasoning, in which the Court
likened workers’ picketing rights to being only for the benefit of “a particular class only”,
while the right to engage in unimpeded commerce was for the benefit of the
“community at large”, and thus of a more fundamental and important nature. Writing
forty years later, Bernie Adell called these statements the “much-reviled” passages in
the decision (Adell, 2003).44 Tucker outlines how this fateful decision was the product of
“the right case, at the right time, with the right people” (Tucker, 2010, at 236).
Another study that also highlights the role of judicial ideology in the historical
development of picketing regulation over time is Sangster’s (2004) study examining the
strike of workers at the Tilco plant in Peterborough in the mid-1960s. Sangster
illustrates how the highly coercive responses of the local judiciary towards picketers
made this strike a cause celebre for the Ontario labour movement, and an important
political precursor to the Royal Commission on Labour Disputes in 1966 that
subsequently led to the adoption of the additional restrictions on injunctions in labour
disputes in section 102 of the Courts of Justice Act.
Gerard and Phillips also highlight the role of judicial ideology and the outcome of
interpretive struggles within the judiciary in the elevation of private property rights over
collective labour rights that occurred in the 1976 decision of the Supreme Court in
Harrison v. Carswell,45 and subsequent caselaw following this as precedent. In this
case, the court held that a shopping mall owner could eject union picketers from a mallowned sidewalk in front of their struck employer’s grocery store, despite the fact that
mall owners had invited members of the public to enter the property. Despite certain
important decisions that distinguished Harrison v. Carswell,46 and some short-lived
legislative reforms (in Ontario), the authors conclude that the majority judgment in this
decision has stood the test of time, and remains a potent source of restriction on
picketing activities that would seek to take place on property owned by third parties,
given modern spatial organization of work and commerce. Gerard and Phillips are also
44
For an early critical response to the Court’s legal reasoning and labour policy effects, see Arthurs (1963).
[1976] 2 S.C.R. 200.
46
See Cadillac Fairview Corp. v. R.W.D.S.U. (1989), 71 O.R. (2d) 206 (C.A.); and R. v. Layton (1986) 33 C.C.C. (3d)
550 (Ont Prov. Ct.).
45
58
Collective Bargaining, Strikes and Lockouts
critical of the legislative reforms adopted in 1992 by the Ontario government that
provided that union members “have the right to be present” on “premises to which the
public normally has access”, even if privately owned, for organizing and picketing
purposes, which required that the property in question had to be “at or near but outside
the entrances and exits to the employees’ workplace.” The authors argued that these
restrictions on the types of property to which picketing workers would have a right of
access was a narrow codification of the ruling in Cadillac Fairview (which distinguished
Harrison v. Carswell), and that they reduced judicial discretion to expand the scope of
this ruling to enable picketer access to other types of locations.
Turning to more recent developments in picketing regulation, MacNeil (2000) similarly
argues that the Supreme Court decisions in Kmart47 and Allsco48 reflect traditional
judicial characterizations of union activity. He criticizes the pains taken to analytically
separate “leafleting” as a protected form of secondary activity distinct from all other
forms of secondary picketing. He argues that the Court’s “invocation of the idea that
conventional picketing, even when peaceful, acts as a signal and a form of social
pressure, rather than as rational and informed discourse, clearly demonstrates the
impact of ideology and metaphor.”49
More recently, the symbolic pronouncement of the Supreme Court in 2002 in RWDSU,
Local 558 v. Pepsi Cola Canada Beverages (West) Ltd50 that picketing is a Charterprotected form of expression, suggests somewhat of a shift in judicial ideology to some
degree,51 but also arguably reinforces the desirability to examine the appropriateness of
this historical foundation of modern picketing regulation (Adell, 2003).
Adell (2003) assesses the impact of the Pepsi case on the state of the law governing
secondary picketing. He argues that a key unresolved question pertaining to the
47
th
(1999), 176 D.L.R. (4 ) 607.
th
(1999), 176 D.L.R. (4 ) 647.
49
Macklem (2000) comes to very similar conclusions in his review of caselaw leading up to and including Kmart and
Allsco.
50
th
(2002), 208 D.L.R. (4 ) 385 (S.C.C.) (hereinafter “Pepsi”).
51
As Tucker suggests, some recent judicial decisions provide some evidence that there has been a shift in the
prevailing judicial image of labour unions from what Arthurs famously quipped as being that of a “churlish
adolescent” requiring additional coercive restraint, to that of “senior citizen” losing the capacity to cope with its
more hostile environment (Tucker, 2010).
48
59
Collective Bargaining, Strikes and Lockouts
“wrongful action model” established by the Court is what sorts of secondary picketing
behaviour, that does not involve a nominate tort or crime, will make secondary picketing
unlawful? He argues that continued reliance upon the body of “industrial torts” is
undesirable, suggesting that they are broadly understood as failing to constitute a
rationally coherent and principled approach to determining illegality, and suggests that
employers should be more strictly required to justify requests for restrictions on
picketing on criminal and/or nominate tort conduct.
Dinsdale and Awrey (2004) provide a comment on the Supreme Court’s decision in
Pepsi. Specifically, they take issue with what they claim is the Court’s reliance upon the
decision of the U.S. Supreme Court in the Tree Fruits52 decision with respect to
secondary picketing. The Tree Fruits decision involves a distinction between two forms
of picketing: general and “struck product” picketing. They suggest that subsequent U.S.
caselaw shows the difficulty of applying these doctrines in the modern economy. Lastly,
they argue that courts underestimate the powerful nature of the “signalling effect” of
picketing, regardless of the stated, literal message.
Smith (2014) provides an in-depth history and analysis of the Pepsi strike in which the
workers challenged the legality of restrictions on their secondary picketing strategies,
and which ultimately led to the Pepsi decision. Smith argues that the history of this
event illustrates that underlying collective solidarity was the most important factor in the
success workers achieved in this strike, generated at least partly through a short-lived
act of collective civil disobedience (plant occupation). It also shows that worker
campaigning and secondary picketing throughout the city was effective in helping to
forge this collective solidarity. A more specific insight is that a key potential benefit of
secondary picketing from the workers’ perspective was that it enabled community-wide
campaigning aimed at reducing the ability of the employer to use replacement workers,
through moral suasion. Smith is also critical of the ultimate judgement in Pepsi, since
despite adopting the wrongful action model, the Supreme Court simultaneously seems
to have legitimized the continued practice of reliance upon the body of excessively
52
National Labor Relations Board v. Fruit and Vegetable Packers & Warehousemen, Local 760 et al., 377 U.S. 58
(1964).
60
Collective Bargaining, Strikes and Lockouts
restrictive industrial torts as the main framework for determining whether picketing
behaviour ought to be found “wrongful”.
Using qualitative empirical methods, De Lint at al. (2005) examine Court and police
practices with respect to picketing contexts and injunction proceedings. Drawing from
certain theoretical literature on “governance”, which emphasizes the decentralized
nature of authority, they highlight how both courts and police systematically engage in
practices of “deferral” of decision-making in labour dispute contexts. Their analysis also
highlights how there has emerged an increasingly normal practice of bargaining in the
shadow of potential intervention (by Courts and police) towards fashioning an agreed
“protocol” pertaining to picketing activity, since there is systematic pressure applied on
parties to do so, including explicit police recommendation and assistance with protocol
formation in many jurisdictions.
Other academic literature has also discussed how, in our regime, there are rather strong
restrictions on worker solidarity beyond individual bargaining unit boundaries (Fudge
and Glasbeek, 1995). One key point of contention is that in our regime, workers’
support for strike action of other workers say, in the form of refusing to cross a picket
line or refusal to carry “struck work”, are highly dangerous ethical choices for workers to
make. Not only does the regime not provide protection for such an ethical choice, but
the law is rather coercive in response to such choices. Supportive behaviour often
leads to the finding, where workers are unionized and multiple individuals are involved,
that this constitutes an illegal strike by the sympathetic workers. Further, sections 83(1)
and 100 of the Act contain broadly worded restrictions on behaviour that might cause
such illegal strikes to occur, often used to restrain secondary picketing activities at
locations involving other unionized workers. Beyond this, the law also imposes severe
limitations on workers ability to negotiate into their own employment conditions
protections for their ability to engage in these kinds of ethical choices in favour of
broader solidarity, since these are generally interpreted as being void against public
policy, insofar as the behaviour often contravenes rules on the timeliness of strikes.
Various legal restrictions on collective action beyond the bounds of the bargaining unit,
or the precise struck location, have been cited in academic research as having a long-
61
Collective Bargaining, Strikes and Lockouts
run corrosive effect on worker solidarity, and aggregate bargaining power (Fudge and
Glasbeek, 1995).
D) Subsequent Interest Arbitration as An Alternative Response to Work
Stoppages
If the pattern of increasingly regular long-duration stoppages in recent years is accepted
as problematic, a key policy response for consideration, well targeted at this precise
phenomenon, would be some sort of interest arbitration mechanism that can be invoked
by one party, after a certain length of work stoppage has already occurred. On the one
hand, this is an “intrusive” measure into the collective bargaining process potentially
raising theoretical concerns about the intangible costs of interest arbitration on
bargaining behaviour. These include concerns about a so-called “narcotic” effect (the
increased dependence on arbitration to settle disputes in future due to loss of
bargaining capacities) and the “chilling” effect (the increased unwillingness to make
concessions in bargaining, based on expectation of usage of interest arbitration and
reluctance to narrow the zone of disagreement in advance of arbitration). On the other
hand, it would seemingly only apply to a small set of disputes, of comparably larger
magnitude. As well, various conditions precedent can also be imposed on the process
for invoking the arbitration mechanism, further limiting its scope of application. Overall,
it seems feasible to design such a process that would still allow for a significant period
of uninterrupted work stoppage, with all its ensuing costs having to be absorbed by
parties for maintaining their bargaining positions, preserving much of the same degree
of economic pressure existing in our current system to reach a settlement. The new
additional threat of a subsequent arbitration could maintain or even spur pressure to
settle in such cases, given the risk of eventual arbitrated outcomes if the stoppage
becomes “excessively” long.
The key issues for consideration in design of such a mechanism are the precise
conditions precedent to be used. In general, the main considerations would appear to
be the choice of required length of the work stoppage; whether there ought to be any
inquiry into the prior bargaining behaviour of the parties up to that time (and whether the
62
Collective Bargaining, Strikes and Lockouts
request ought to be linked with some “good faith” requirement); and whether there ought
to be any additional inquiry into how likely the parties are to reach a settlement in some
measurable future time, without arbitration. A subsequent-arbitration mechanism was
adopted in Manitoba in 2000,53 with certain further clarifying amendments adopted in
2004.54 To date, there does not appear to have been any academic analysis of any sort
relating to these provisions, nor any other writing about such a provision in general,
aside from some very brief descriptive comments about these specific Manitoba
provisions in David Doorey’s online “blog” (Doorey, 2012) and a similarly brief comment
about them in a newspaper editorial written by Unifor economist Jim Stanford (Stanford,
2012). What follows is therefore not a literature review about this model (as originally
requested by the CWR), but a very brief outline of the Manitoba mechanism, and a
summary of the limited information available about how this mechanism has functioned
since 2000. The exact provisions of the Manitoba statute establishing this mechanism
are contained in Appendix 3.
The basic requirements of the Manitoba mechanism are as follows. First, the work
stoppage must have lasted for at least 60 days already prior to any application for
arbitration being filed. As a further condition, the parties must have attempted to reach
an agreement, with the assistance of a conciliation officer, for at least 30 days during
the stoppage. After receiving the application, the Board must inquire into whether good
faith bargaining has occurred and whether the parties are likely to reach an agreement
within 30 days if they continue bargaining, and must make these determinations within
21 days, although the Board may delay making these determinations to the point where
it is satisfied that the applicant party has bargained “sufficiently and seriously” with
respect to the matters still in dispute. If the Board finds both good faith bargaining and a
likelihood of settlement within 30 days, the Board will refuse to order arbitration, but may
order further conciliation or mediation.
If the Board finds a lack of good faith
bargaining by the applicant, the applicant may still file another application for arbitration
at a later date. If there is still no agreement after 30 days expire, either party may file a
new application for arbitration. Once the Board decides arbitration is appropriate, there
53
54
The Labour Relations Amendment Act (2), S.M. 2000, c.45.
The Labour Relations Amendment Act, S.M. 2004, c. 31.
63
Collective Bargaining, Strikes and Lockouts
must be an immediate cessation of the work stoppage and a return to work, and the
parties also have 10 days to agree to use interest arbitration and to submit the name of
the arbitrator they have agreed upon. If the parties do not agree to arbitration and a
specific arbitrator, then the new agreement will be settled by the Board itself. The new
agreement is effective for a period of 1 year following the expiry date of the previous
agreement, or for any longer period agreed to by the parties. If the agreement is
settled more than 6 months past the expiry of the previous contract, the new agreement
shall remain in effect for 6 months only.
Overall, it is apparent from a review of the provisions that a number of checks have
been imposed to try to limit the recourse to the provision, and the arbitrated outcomes
only last for rather short periods of time. The evidence available about the function of
these provisions is very limited. What we do know is that since they were adopted in
2000, there have only been a total of seven (7) applications ever filed under these
provisions, at a minimum suggesting that there is little evidence of any “floodgate” of
arbitration activity that has been opened by adopting these rules. Further, three (3) of
these applications were voluntarily withdrawn, perhaps related to the perceived
outcome of these applications. Thus, in fifteen (15) years, the Board has only been
required to make four (4) rulings on applications under this provision,55 and the Board
granted each of these four applications.
One provision in the statute requires that the Manitoba Labour Management Review
Committee to report every two years on the functioning of this provision. A review of the
reports of these committees since 2000 shows that the feedback on these provisions
has been extremely minimal. The committee offered no feedback until 2009, apparently
because it was occupied with more pressing and substantial matters. This 2009 report
contained a joint recommendation that in contexts where settlement occurs more than 6
months after the expiry of the last agreement, arbitrated outcomes should be extended
by more than the rather short 6 month extension provided by the Act. The Committee
recommended that a one year extension of the arbitrated agreement would be more
55
See 2007 Canlii81873; 2007 Canlii81878; 2010 Canlii99140; and 2011 Canlii98421. These decisions are also
available from the author upon request.
64
Collective Bargaining, Strikes and Lockouts
appropriate. The Committee’s subsequent 2011 report made no recommendations for
change, merely reiterating that the various conditions precedent to invoking the
provision ought to remain in the Act. Lastly, the Committee’s 2013 report merely noted
that additional utilization of existing mediation services would be helpful in resolving
disputes. So, overall, the main feedback provided on the functioning of this rarely
invoked mechanism has been a recommendation favouring slightly longer arbitrated
settlements to prevent the occurrence of very short agreements in unique cases.
In its review of work stoppages in the federal sector, the Annis Report (Annis, 2008)
observed similar patterns in work stoppage dynamics in the federal sector as those we
have witnessed in Ontario (notably, fewer stoppages, of longer average duration). The
Report included this sort of subsequent interest arbitration in its review of some
alternative responses to stoppages. The Report notes that in order to further reduce
any potential “chill” on bargaining that might arise from expected access to this
provision, it is possible to consider adding more uncertainty as to the availability of the
mechanism, such as some form of Ministerial discretion. The Report also agrees that
the new threat of an unfavourable arbitrated settlement could also encourage
settlement, countering concerns about the “chilling” of bargaining (Annis, 2008, at 75).
The Annis Report also identified some other potential responses to work stoppages,
most of which have already been discussed in this Report. One additional alternative it
cited was the possibility of a mandatory return to work “cooling off period” in stalemated
work stoppages, combined with mediation assistance. The theory here is that in some
cases of stalemate, a temporary return to work might somehow catalyze some existing
opportunity for settlement. Interestingly, both labour and management stakeholders
expressed opposition to this sort of mechanism in their feedback to the Commission.
Lastly, it should be noted that under current Ontario law, no such access to an
arbitrated settlement is available in the broader private sector, without both parties’
consent, except for limited access to first contract arbitration, which is beyond the scope
of this report. This is true even in cases where the OLRB finds that a party has not
bargained in good faith. Although 1993 amendments to the Act granted the OLRB the
authority to order arbitration as a remedy as a response to bad faith bargaining, this
65
Collective Bargaining, Strikes and Lockouts
provision was repealed in 1995. Literature on the duty to bargain in good faith is
reviewed in the next section.
E) The Duty to Bargain in Good Faith (“DBGF”) - section 17 of the LRA
Section 17 of the LRA contains the “duty to bargain in good faith” (DBGF). Somewhat
of a two-pronged measure, it obliges both parties to “bargain in good faith and make
every reasonable effort to make a collective agreement.” This is a longstanding
provision in Ontario law. Since 1948, Ontario law has contained an explicit DBGF
requirement with very similar terminology.56 The rule was to a certain extent a
transplantation of the similarly named duty enshrined in the U.S. National Labour
Relations Act (“NLRA”).
In earlier years, the DBGF was enforced by quasi-criminal enforcement provisions. This
approach to regulation of bargaining behaviour became subject to significant criticism,
in that it seemingly placed significant limits on the scope of supervised behaviour and
the range of remedies to address undesirable behaviour, resulting in limited litigation
and application of the DBGF mechanism (Palmer, 1969; Adell, 1980). The 1975
amendments to the Act addressed these concerns squarely, and “breathed new life into
the duty” to some extent (Bendel, 1980). In place of the previous limitation of OLRB
power primarily to determining whether quasi-criminal prosecutions could proceed, the
new provisions empowered and directed the OLRB to investigate alleged violations of
the duty and to “determine what if anything, [the respondent] shall do or refrain from
doing with respect thereto”.57 Subsequently, there was a large volume of litigation and
caselaw development fleshing out the scope, parameters, and content of the DBGF.
In a classic 1958 article, Cox (1958) outlined four key policy goals underlying the DBGF
in the U.S.: 1) prevention of recognition strikes; 2) prevention of union avoidance
through non-recognition, thereby supporting growth of strong labour unions; 3) support
for collective (as opposed to individual) bargaining; and 4) the promotion of collective
56
Ontario Labour Relations Act, 1948, s 12. The 1950 version of the statute introduced more explicit language
pertaining to a “good faith” requirement in bargaining. Section 11 of this statute required parties to “bargain in
good faith and make every reasonable effort to make a collective agreement”. The DBGF has continued to be
expressed in these terms in Ontario since 1950. See Bendel, 1980, p.1.
57
S.O. 1975, c. 76, s. 21(1).
66
Collective Bargaining, Strikes and Lockouts
bargaining as a “rational process of persuasion.” The 4th goal moves beyond merely
supporting the shift from an individual to a collective bargaining regime, it involves
altering the bargaining behaviour of the parties, where this is deemed objectively
desirable. This is therefore an additional encroachment upon the notion of a pure
“freedom of contract” than necessitated by the other goals, and may be understood as
balancing the goal of freedom of contract with other goals, such as the promotion of
industrial peace. Bendel (1980) found support for all four of these policy goals in early
Ontario caselaw, and comparably greater support for the 4th goal than occurs in the U.S.
One reason for this may be based in the wording of the NLRA itself, since the 1947
Taft-Hartly amendments to the NLRA reinforced the principle of freedom of contract, in
specifying that the DBGF does not “compel either party to agree to a proposal or require
the making of a concession”.58 On the other hand, although Canadian tribunals like the
OLRB have generally been comparably more willing to supervise undesirable
bargaining behaviour than U.S. adjudicators, Bendel argues that certain adjudicative
trends, such as the reluctance to impose a collective agreement as a remedy, and the
reluctance to draw an inference of bad faith bargaining solely from the substantive
bargaining proposals of one of the parties, suggests that Canadian adjudicators have
nevertheless imported the underlying U.S. bias in favour of freedom of contract over
other goals to a significant degree.59
One of the well-known distinctions between Canadian and U.S. jurisprudence over the
DBGF is the U.S. practice of categorizing all potential bargaining terms as failing within
either a “mandatory” or “voluntary” category of terms. It is an unfair labour practice in
the U.S. for a party to “insist” upon (bargain to impasse, engage in industrial action) its
position on a “voluntary” item. Again, this practice may be related to the existence of
comparably more restrictive language in the NLRA specifying that the range of terms
subject to the DBGF are “wages, hours and other terms and conditions of
employment”60 and U.S. adjudicators have sought to draw a line between demands that
fall within and outside this category. As Langille (1983) notes, this approach has been
58
Labor-Management Relations (Taft-Hartley) Act, (LMRA) 29 U.S.C.S. ss. 151-67 (1975).
Bendel, p 14-15.
60
NLRA, s. 8(d).
59
67
Collective Bargaining, Strikes and Lockouts
“widely criticised on a number of functional and instrumental bases.” One key critique of
this approach relates to its effect, insofar as it led to the juridical construction of a range
of protected management prerogatives. Various issues relating to organizational
strategy and design, including restrictions on subcontracting of work operations, were
held to be “voluntary” issues (Langille, 1983). In hindsight, these legal developments
likely contributed significantly to the subsequent weakness and inability of U.S. unions
to resist the spread of employer reorganization strategies that expanded rapidly in the
1980’s and 1990’s, and which contributed significantly to the decline in unionization in
the U.S. (Kochan, Katz and McKersie, 1986). While one may interpret these sort of
legal restrictions as merely being an extension of the U.S. preference for freedom of
contract (more precisely, protecting the employer’s freedom not to contract over certain
terms) one may alternatively view the law here as actively constraining union bargaining
freedom.
While Canadian tribunals did not explicitly follow the more blatant “mandatory/voluntary”
approach, academic critique points to analogous forms of managerial prerogatives that
were effectively imposed into our law. Langille (1983) notes that although Canadian law
formally allows bargaining over subcontracting and other key organizational decisions,
there has been a refusal to promote the principle of “partnership” with regard to these
key decisions to any serious extent. He notes that outside the precise statutory
timetable for bargaining, the law preserves unilateral managerial authority over these
matters at all other times. He argues that while the law provides that significant
organizational plans (which would be within the scope of the duty to bargain) and/or
their contemplation may be required to be disclosed depending upon their timing, this
regulatory approach overall simply establishes managerial incentives for manipulation of
the process and timing of the contemplation and finalization of such strategic plans in
order to avoid having these issues become subject to collective bargaining (Langille,
1983). This ability of employers to prevent key organizational strategic decisions from
becoming subject to collective bargaining was further supported by earlier
developments in arbitration caselaw concerning contracting out. While there was
initially two competing views over whether employers retained an underlying prerogative
to engage in contracting out of bargaining unit work (subject to restrictions in the
68
Collective Bargaining, Strikes and Lockouts
collective agreement) the view supporting the managerial prerogative eventually “won
the day”.61 As Langille (1983) notes, this outcome in the arbitration case law had a
spillover effect, in turn eroding any “partnership” based requirement that could be
imposed by the DBGF.
As well, somewhat analogous to the U.S. approach, the DBGF imposes restrictions on
“bargaining to impasse” various forms of demands deemed (ex post) “illegal”. Carter
argued that Canadian tribunals have been willing “to define the concept of illegality in its
broadest sense” (Carter, 1983). Restrictions in this “broadest sense” include restrictions
of demands that are not overtly illegal such as a clear violation of a legislated minimum
employment standard, but those that are deemed inconsistent with the overall
framework of the statutory regime, such as attempts to negotiate changes to the
existing scope of the bargaining unit, or insistence upon an alternative, broader-based
bargaining structure (Carter, 1983). The issue of union ability to leverage its existing
bargaining power to pursue broader based bargaining would seem to be an increasingly
important (even if not salient) policy question, given organizational fissuring in the
modern economy (Fudge and Glasbeek, 1995).
In light of the goals of the DBGF, an acutely well-known source of tension in the law is
the distinction between “hard” bargaining, deemed acceptable, and so-called “surface”
bargaining, based on a desire to avoid reaching a collective agreement, and/or to
undermine the union as the bargaining agent. This analytical distinction has come
under scrutiny. Langille and Macklem (1988) argue that DBGF jurisprudence is built on
a “contractualist” philosophical foundation, with primary reliance upon a procedural,
rather than substantive conception of justice. In formal terms, the law embraces the
employer’s ability to use its superior bargaining power to maximize its self-interest, while
guarding only against behaviour aimed at undermining collective bargaining per se.
Langille and Macklem (1988) note that a key weakness of this approach is rooted in the
reliance upon identifying subjective employer intention to avoid collective bargaining,
and the implicit default acceptance within the jurisprudence that employer self-interest
61
See in particular the key decision in Russelsteel Ltd., (1966), 17 L.A.C. 253.
69
Collective Bargaining, Strikes and Lockouts
maximizing behaviour is legitimate or “rational”. This leads to the consequence that the
DBGF
“can capture those employers who view it in their self-interest not to recognize
the trade union. If your intent is not to sign a collective agreement, because this
is your view of how to maximize your self-interest then you are, if the evidentiary
problem is overcome, caught by the duty. However, if you are willing to sign a
collective agreement, then you can maximize your self-interest. The problem
with this neat view of the world is that it captures only: a) the rational but less
powerful and b) the irrational but more powerful, employers of the world. It does
not capture the rational and powerful employer.”
Langille (1983) similarly captures the weakness in an approach that assumes the ability
to distinguish and disentangle economically rational self-interested behaviour from
illegitimate self-interested behaviour:
“can one say that a motive is an economic one when the economic impact
referred to is the ‘axiomatic’ economic impact of unionization?...It would be an
exercise in supreme sophistry if it could be alleged that the motive was not antiunion because the objection taken was not to the union itself, but to its
(‘axiomatic’) effects.”
In addition, it is reasonable to conclude that the emphasis on subjective intention
creates considerable difficulties of proof in the enforcement of the DBGF, and this is
commonly cited as an additional weakness of the DBGF mechanism. As a solution to
these weaknesses, Langille and Macklem (1988) propose that the law ought to
recognize a substantive content to the DBGF, and that there ought to be imposed a
clearer obligation to negotiate a collective agreement containing particular terms (they
suggest in particular provisions dealing with seniority, just cause dismissal protection,
and fair and accessible grievance procedures).
Overall then, in theoretical terms, the DBGF has not been understood as being aimed at
performing a significant power-brokering function as between the parties, aside from its
function of protecting the shift to collectivism and preserving the bargaining structure
70
Collective Bargaining, Strikes and Lockouts
established by other instruments within the regime (Cox’s goals # 1, 2 and 3). Other
than this protective role, the DBGF is understood as being aimed primarily at increasing
the ability of the parties to otherwise freely pursue their own62 self-interest, given their
existing degree of power. The only caveat here is where the DBGF may address
behaviour that would systematically alter relative bargaining power, such as
asymmetrical information pertaining to the employer enterprise. This conception of the
DBGF seems based on the assumption that other aspects of the regime sufficiently
carry the burden of addressing the question of the balance of power.
A final comment may be made about how the DBGF fits within the overall statutory
regime. It is axiomatic that the DBGF only applies post-certification, and thus has no
application whatsoever to employers whose workers have not yet proceeded
successfully through the certification procedure63and given prevailing trends, this is the
vast majority of (private sector) employers. Thus, most employer behaviour in the
marketplace is free from any form of “good faith” supervision, even with respect to
behaviour that may be designed to avoid collective bargaining. A key consequence in
the modern labour market is that many organizations are free to adopt “fissured”
structures, relying upon forms of subcontracting, franchising, and supply chains in
organizing work. As one potential policy response to this new reality, Noah Zatz has
argued that unfair labour practice law ought to be expanded to increasingly supervise
pre-certification employer behaviour, particularly strategic decisions with respect to the
adoption of particular organizational structures that may be driven by union avoidance
(Zatz, 2010). While it is not clear whether we would conceive of such new ULP content
as being an expansion of the DBGF, or rather as falling under a different content label, it
is mentioned here insofar as it relates to the issue of potential regulatory supervision
over organizational structures that may be either aimed at, or have the effect of,
undermining collective bargaining.
62
This leaves aside complications in the accumulation of “preferences” of large numbers of bargaining unit
employees, and in assuming a “self”-interest to employer organizations.
63
Or some lawful form of voluntary recognition, where available.
71
Collective Bargaining, Strikes and Lockouts
F) The grievance arbitration system
The literature has cited numerous concerns about the functioning of the grievance
arbitration system. Some of the most significant concerns identified, which may be
inter-related, are delay, increasing legalism, cost, and the scope of arbitral jurisdiction.
It is interesting to note that various modern concerns about the grievance arbitration
system are not new. As early as the 1970’s various commentators and parties in labour
relations were lamenting the growth in excess legalism and the ensuing delay that
seemed to have already set into grievance arbitration by that time (Beatty, 1974). Thus,
the long trajectory of studies overall suggest that what we appear to have experienced
in recent decades with respect to delay, cost and legalism, is an increasingly worse
“new normal” over time.
The Ontario government responded to early concerns about growth in delay and
legalism in grievance arbitration in 1979 with a significant policy reform: “expedited
arbitration”. Rose (1986) reviews the extent to which the new system was adopted by
parties and its effectiveness in reducing delay. Rose concludes that by several
measures, the policy reform was a success. The new procedures were voluntarily
adopted in large measure by parties, encouraged more settlement by mediation, and
reduced the time of grievance disposition substantially. However, it is acknowledged
that the new process only partially addressed the various concerns about legalism and
delay, since it did not go as far as to provide for limitations on the arbitral process itself,
but rather focused on some factors (Ie arbitrator appointment) determining delay in the
pre-hearing stage. Rose also found evidence that in cases involving discharge of the
employee, delay seemed to be correlated with a reduced likelihood that the arbitrator
would find in favour of the grievor and grant reinstatement.
Ponak and Olson (1992) examined Alberta grievance arbitration decisions from 1985-88
to determine trends in delay and some of the factors causing delay. The analysis
examines delay in the pre-hearing and post-hearing stages, and suggests that in the
data reviewed, most delay is in the pre-hearing stage. Statistics on average delay
across aggregate caseloads however are distorted by the fact that some grievances
72
Collective Bargaining, Strikes and Lockouts
take an inordinate length of time, so that median delay is substantially less than mean
delay. Factors that tend to be positively correlated with increased delay include the use
of 3 person panels, the sector (a mild public sector effect); and the type of issue. The
analysis suggested that refinements in the arbitrator selection and scheduling stages in
Alberta would help reduce delay.
Thornicroft (1995) reviews Newfoundland arbitration outcomes from 1980-1992, and
also examines trends and factors involved in delay, and also distinguishes delay in the
pre-hearing from the post-hearing stages. Factors involved in delay included the use of
3 person panels; the use of lawyers; the choice of individual arbitrator; and a mild
public-sector effect.
Ponak et al. (1996) also examined grievance delay based on a review of Alberta
arbitration decisions from 1985-88, separating out 4 stages in analysis: the prearbitration grievance stage, the arbitrator selection stage, the hearing scheduling stage,
and the preparation of the arbitration award stage (I.e., post-hearing). They argue that
their findings suggest that the causes of delay are different at each stage.64
They
found that discharge cases are handled relatively faster. The presence of legal counsel
increases delay during the scheduling stage, but reduces delay in the arbitrator
selection and decision preparation stages. There was little evidence to support the
view that arbitrator workload caused greater delay.
Senior members of the bar and judiciary have also lately commented on the plight of the
grievance arbitration system. Whitaker (2010) reviews the historical development of
large scale adoption of mediation-arbitration practices in Ontario. The key flashpoint
was the development of huge backlogs of grievances by the late 1980s and early 1990s
within various large scale bargaining structures in the province, and a growing
recognition of the necessity to use mediation on a larger scale to reduce caseloads and
their burden. Gradually, most arbitrators in Ontario became expected to be able to
64
They find primarily that the length of time from the filing off a grievance to referral to arbitration is a function of
the complexity and type of issue; delay in arbitrator selection is associated with the use of legal counsel and the
size of the arbitration board; scheduling delay is associated with the nature of the grievance and the use of outside
legal counsel; and delay in preparing the decision is linked to the complexity and type of the issue, board size, the
presence of legal counsel, and the arbitrator’s workload.
73
Collective Bargaining, Strikes and Lockouts
provide med-arb if the parties desired; it has become a normal and essential part of the
toolkit.
Winkler C.J.O. (2010, 2011) contrasts modern labour arbitration practices with those of
an earlier “golden age” of arbitration. The demise of the “golden age” has been driven
by factors including expanded arbitral jurisdiction and a cultural shift within arbitration
from pragmatism to legalism. On the point of expanded arbitral jurisdiction, he cites the
outcomes in Weber v. Ontario Hydro65 and in Parry Sound v. OPSEU66 as crucial, since
they increased the flow of disputes to arbitration rather than other processes. Winkler
argues that overall the grievance arbitration system has lost “proportionality” and that
solutions necessarily involve the re-insertion of a new ethic of proportionality into the
system. That is, there needs to be a new ethic that various litigation practices ought to
be engaged in only if they are proportionate to the importance of the matters at issue in
the grievance. From the text of his two public presentations, Winkler seems to couch
his suggestions as primarily being aimed at the parties themselves. However, his
insights may also be interpreted as supporting policy intervention, particularly if the
parties, or perhaps their legal counsel, face systemic incentives generating nonproportional legalism, and thus have limited ability to voluntarily opt out of this
behaviour. Similarly, Pink and Wallbridge (2010) decry the excess legalism that
lawyers have imposed on the process over time, stating that “lawyers have made the
process so complicated and expensive that we have ‘killed the goose that has made our
careers’’”.
It should be noted that the expansion of arbitral jurisdiction has been driven by multiple
developments, and has very important implications for any assessment of the
functionality of the grievance/arbitration system. Not only has the scope of arbitration
been expanded by judicial decisions in Weber and Parry Sound, but also by statutory
provisions over time that directed the flow of employment standards and human rights
disputes into arbitration. While this expanded jurisdiction may have produced
65
(1995), 2 S.C.R. 929. The Court ruled that the matter in dispute arose out of the collective agreement, and that
arbitrators had exclusive jurisdiction to determine such matters.
66
Parry Sound (District) Social Services Administration Board v. Ontario Public Services Employees Union, Local 324,
[2003] 2 S.C.R. 157. Here the Court expanded the principle in Weber in effectively holding that arbitrators had
jurisdiction to apply employment-related statutes. expanding the volume of
74
Collective Bargaining, Strikes and Lockouts
additional costs and/or delay within the system, there may also have been some
enhancement in adjudicative rationality and/or efficiency under the consolidation of
employment-related adjudicative authority in the arbitral forum. In this vein, Shilton
(2013) argues that arbitration should be confirmed as the exclusive forum for human
rights claims of unionized workers that are closely linked to the collective agreement,
and that recourse to statutory human rights tribunals should be reserved only to unique
cases, such as those where union complicity in discrimination is in issue.
Curran (2015) provides the most up-to-date analysis of grievance arbitration delay in
Ontario. He provides summary statistics on various trends associated with delay in
arbitration, and also estimates of the effects of separate factors on grievance delay in
each of the pre-hearing, hearing and post-hearing stages. Compiling the results of
previous studies with respect to arbitration delay, he shows that they collectively reveal
a continuous upward trend in arbitral delay, with an estimated average annual growth of
approximately 11.9 days per year, from the mid-1970s to the mid-1990s. Then, using a
sample of 15% of all arbitrated grievance resolutions in each of 1994, 2004, and 2012,
he compares mean values for a number of interesting variables. For example, he
shows that average total disposition time grew from 419.9 days in 1994, to 444.12 days
in 2004, and to 589 days in 2012. Thus, the very high growth in grievance duration in
Ontario in recent decades comes after decades of substantial growth in grievance
duration already occurred. It should be noted that the design of most of these studies
only examines the subset of grievances that proceed to arbitration and result in an
award being issued. So, comparisons are capturing differences over time within this
smaller subset of grievances filed, not the average behaviour of all grievances filed.
Curran (2015) shows that most of the recent growth in overall grievance delay in this
subset has been concentrated in the pre-hearing and hearing phases. Some other
trends stood out in his comparisons of means, pertaining to increased legalism and/or
arbitral jurisdiction. The average number of cases cited in each arbitration award grew
substantially from 5 in 1994 to 7.3 in 2012. The filing of legal briefs increased
substantially, occurring in 8% of grievances in 1994, and in 23% of grievance in 2012.
Average decision length grew from 35 paragraphs in 1994 to 43 in 2012. As well, the
75
Collective Bargaining, Strikes and Lockouts
average workload of arbitrators also grew, with an average workload of 16.7 cases in
1994 to 22.5 in 2012.
Curran (2015) also performs more complicated quantitative analysis, using Cox
proportional hazard models, designed to test how a number of more specific factors
(variables) correlate with grievance duration, in his sample of 397 observations.
These
models provided little evidence for the proposition that delay was being driven by
increased arbitrator workload. However, evidence did support the importance of
growing legalism. The use of lawyers was associated with increased duration in both
the pre-hearing and hearing phase. Preliminary objections were associated with longer
pre-hearing phases. Additional witnesses seemed to lengthen the hearing phase, and
total disposition time, and expert witnesses were associated with a longer hearing
phase. Legal briefs were associated with lengthier decision preparation phase.
Credibility challenges delayed the decision preparation phase and total disposition time.
Interim awards lengthened the hearing phase. Estoppel arguments lengthened the
hearing phase. On the other hand, the citing of “labour relations considerations” was
associated with reduced delay in the prehearing phase.
With respect to the effect of
expanded arbitral jurisdiction, his study found conflicting results. For example,
grievances involving the ESA had significantly longer pre-hearing phases, and human
rights cases had longer hearing and post-hearing phases. On the other hand, after the
Parry Sound decision, grievances involving human rights were shorter in the prehearing stage, and were not lengthier overall. Interestingly, he also found that the use
of med-arb lengthened total disposition time, although this finding only applies with
respect to grievances that ultimately require arbitration awards because they are not
resolved by mediation.
In considering the overall relevance of the various academic studies of the grievance
arbitration system, it is important to reiterate that the data points used in these studies
are completed arbitration awards, and thus the delay being measured pertains only to
the subset of grievances that actually proceed to an arbitral award. This is important
given that most grievances settle prior to arbitration, or at arbitration. Thornicroft (2009)
estimates that only about 2-3% of all grievances filed proceed all the way to an arbitral
76
Collective Bargaining, Strikes and Lockouts
award. Further, there is an anecdotal impression noted by many in the field that an
extremely high percentage of grievances that proceed to arbitration actually settle at the
first day of an arbitration hearing. These settlements at arbitration are not included in
the data in these studies, there being no reliable aggregate dataset constructed yet that
contains these events. This raises the question of whether increasing grievance delay
captured in these studies is a matter of selection bias. Are the cases that proceed to
arbitration of a select type or degree of complexity, increasing both the necessity to
proceed to an arbitral award and greater time for disposition? We do not have a
definitive answer to this question. However, there are grounds for the view that the
evidence amassed in these studies suggesting increasing grievance delay over time
ought not to be dismissed by this selection bias theory. For example, much of the total
delay effect due to legal complexity would presumably register in the hearing and/or
decision preparation phase. However, it is notable that from 1994 to 2012, average
delay in the pre-hearing phase also grew from 273 to 397 days (Curran, 2015). While it
is conceivable that there is some selection bias causing the observations of increasing
grievance delay, many of these estimates still constitute prima facie evidence of
significant growth in average grievance delay in the system.
Overall, there is fairly clear evidence that the time it takes for a grievance to progress
from filing to an ultimate award has been growing steadily in Ontario for decades, with
recent evidence supporting a significant growth trend up to as recent as 2012 (Curran,
2015). Certain findings support the view that excess legalism and arbitral jurisdiction
are driving this trend, although evidence seems to suggest that the processes are
somewhat nuanced in their effects upon different stages of the process. These findings
along with increasingly common statements from senior members of the bar and/or
judiciary about the system suggest that policy reforms aimed at reducing delay, if they
were mindful of the different stages of the process, may be increasingly warranted.
Such reforms would seemingly benefit from addressing the lack of proportionality in the
system cited by C.J.O. Winkler, and from paying careful attention to the systemic legal
constraints and incentives facing parties within the arbitration system altering their
77
Collective Bargaining, Strikes and Lockouts
ability to voluntarily opt into more proportionate practices.
78
Collective Bargaining, Strikes and Lockouts
Conclusions and Options
This Report has provided a review of empirical trends in union wage premiums and in
work stoppages activity, and a review of various bodies of literature examining multiple
aspects of the legal regime governing collective bargaining in Ontario. Throughout the
report, some insights provided by these empirical trends, and by academic analysis of
the collective bargaining regime, have been discussed. Rather than reiterating the
various insights, a few additional concluding comments are offered.
In general, the two main sets of empirical trends reviewed, union wage effects and
strike dynamics, are capable of supporting a similar overall interpretation. That is, they
both point theoretically to a major loss of worker/union bargaining power over time in
recent decades. The loss of union wage premiums is an output measurement of this
erosion. The decline in strikes is an additional indicator, and arguably the most
reasonable interpretation based on the evidence overall (including the concurring
erosion of union premiums) is that various socio-economic contextual shifts (discussed
in the report) have eroded the instrumentality of strikes and union bargaining power, in a
simultaneous and reciprocal manner. The erosion of worker collective power is the
combined product of these contextual shifts occurring, against the backdrop of a labour
law regime that we have not adjusted towards counter-balancing these developments.
Other labour law developments over time, discussed in the literature review throughout
the report, have at times contributed to the general erosion of bargaining power, while
certain other supportive developments may be understood rather as being insufficient to
counter this core trend.
Of course, it is possible that policy choices are greatly influenced by conflicting values
about the desirability of reinforcing worker collective action and bargaining power, or
perhaps about the desirability of reinforcing worker collective action and bargaining
power if this means tolerating the occurrence and/or consequences of comparably more
strikes than the few that seem to occur in the modern era. Indeed, it is also possible that
labour policy-makers may be, for whatever reason, somewhat more inclined to look
79
Collective Bargaining, Strikes and Lockouts
primarily to other policy instruments for our primary responses to developments such as
growth in precarity in work and employment, which do not directly involve addressing
the erosion in worker collective action and bargaining power. Conceivably, economic
insecurity may be addressed via other policy domains, beyond what we have
traditionally characterize as “labour and employment law” and may involve forms of
expanded state provision of social welfare. Further, within the domain of labour and
employment law, some may conceive of the option of expanding various forms of
minimum employment standards as a form of substitute for the generally diminished
“union effect” in the labour market. Assuming that it can be said that there has been a
policy response aimed at ameliorating precarious employment67 this latter approach of
emphasizing certain employment standards instruments, focused at the margin of the
labour market, has arguably been the dominant policy response in recent decades
(Vosko, 2010). The fundamental problem with this approach is that it does extremely
little, if anything, to ameliorate shifts in underlying relational power pervading the labour
market broadly, resulting in minimal sustainable effects.
Labour policy overall demands at a minimum an honest recognition of the decline in
collective worker bargaining power and the widespread consequences of this
development, even if only for accurately assessing the nature and acceptability of all
other forms of policy towards inequality and employment/work precarity in work and
employment, absent labour law reform aimed at the ameliorating the decline of worker
collective bargaining power. Indeed, some of the analysis of union effects that continue
to remain in the labour market, outlined in the beginning of this paper (Gomez and
Lamb, 2015) highlight the potential of collective bargaining power, where it can be
constructed under our regime, to produce very significant improvements in employment
conditions and reduced precarity, even where non-standard employment relations are
maintained, suggesting that enhanced bargaining power may ultimately be a more
effective sterilant against precarity than minimum standards instruments.68
67
Indeed, it is quite possible to view the trajectory of labour policy overall, including the abandonment of
collective bargaining as an instrument, as being a direct determinant of increasing precarity.
68
There is also plenty of anecdotal evidence supporting this view. Outcomes of a recently concluded round of
collective bargaining in the grocery retail sector in Ontario provide an eye-opening example. By consciously
80
Collective Bargaining, Strikes and Lockouts
Lastly, this author was also asked to comment about potential concerns in relation to
the availability of data to study the phenomena discussed in this report. The following
comments are offered in this regard.
It would appear that although this was once historically the purview of the Ministry of
Labour, nobody collects information on the actual usage of replacement workers in
collective bargaining disputes in Ontario, and this may also be the case in other
provinces. Since this has been cited as an important potential strategic response to
work stoppages, and since it may be helpful to know how often and in what manners
they are used, this seems unfortunate. It seems as though it would be helpful to know
more information about this phenomenon, such as the volumes of replacement workers
used, the characteristics of the users, and the different “sources” of these workers used
(are they internal managers, personnel transferred from elsewhere in the organization,
new short term “hires”, or agency-supplied workers).
Further, there is also limited data collection performed about aggregate picketing activity
and restrictions on picketing activity, such as the various injunctions and/or picketing
protocols drafted, with or without police assistance, overall.
Also, data on the distribution of unionization in general is somewhat limited. Improved
measures of the union density by industry and sector, among other characteristics might
be helpful in examining various questions about the functioning of the collective
bargaining regime. For example, data on the underlying distribution of unionized
employment by bargaining unit size, was not available.
Lastly, while studies of arbitration have shown the increased burden of this system in
terms of time delay, measures of the burden of developments in the system overall in
terms of costs to the parties seem comparably less available.
prioritizing those with the worst wage, hours, and scheduling conditions (many of which were defined by legislated
minima) under an explicit theme of “raising the floor”, Unifor Local 414 recently obtained some extraordinary
gains for marginal workers. See Mojtehedzadeh (2015). Details of this negotiation were outlined in a presentation
by Angelo DiCaro at the Canada-China Labour Forum, Ryerson University, November 18, 2015.
81
Collective Bargaining, Strikes and Lockouts
References
Adell, Bernard. “The Duty to Bargain in Good Faith: Its Recent Development in Canada.”
Kingston: Industrial Relations Centre, Queen’s University, 1980.
Adell, Bernard. “Secondary Picketing after Pepsi-Cola: What’s Clear, and What Isn’t?” Canadian
Labour and Employment Law Journal. 10 (2003): p. 135-159.
Annis, Peter. Work Stoppages in the Federal Private Sector: Innovative Solutions. Ottawa:
Human Resources and Skills Development Canada, 2008. Retrieved:
Arthurs, Harry. “Labour Law – Secondary Picketing – Per Se Illegality – Public Policy.”
Canadian Bar Review. 41 (1963): p. 573-586.
Ashenfelter, Orley and George Johnson. “Bargaining Theory, Trade Unions, and Industrial
Strike Activity.” American Economic Review. 59 (1969): p.35-49.
Bartkiw, Timothy. “Manufacturing Descent? Labour Law and Union Organizing in Ontario.”
Canadian Public Policy. 34.1 (2008): p. 111-131.
Bartkiw, Timothy. “Baby Steps? Towards the Regulation of Temporary Help Agency
Employment in Canada.” Comparative Labor Law and Policy Journal. 31.1 (2009): p. 163-206.
Bartkiw, Timothy. “Unions and Temporary Help Agency Employment in Canada.” Relations
industrielles/Industrial Relations. 67.3 (2012): p. 453-476.
Bartkiw, Timothy. “Labour Law and Triangular Employment Growth: A Theory of Regulatory
Differentials.” International Journal of Comparative Labour Law and Industrial Relations. 30.4
(2014): p. 413-433.
Bartkiw, Timothy. “Review of The Fissured Workplace: Why Work Became So Bad for So Many
and What Can Be Done to Improve It”, by David Weil.” Comparative Labor Law and Policy
Journal. 36.2 (2015): p. 323-331.
Beatty, David. “Procedural Irregularities in Grievance Arbitration.” McGill Law Journal. 20.3
(1974): p. 378-393.
Bendel, Michael. “A Rational Process of Persuasion: Good Faith Bargaining in Ontario.” The
University of Toronto Law Journal. 30.1 (1980): p.1-45.
Benjamin, Dwayne, Morley Gunderson, Thomas Lemieux, and Craig Riddell. Labour Market
Economics, 7th ed. (Toronto: McGraw Hill, 2012).
82
Collective Bargaining, Strikes and Lockouts
Brym, Robert. “Affluence, Power, and Strikes in Canada, 1973-2005”. Social Inequality in
Canada: Patterns, Problems and Policies, 5th edition. Edited by Edward Grabb and Neil Guppy
Toronto: Pearson, 2009, p. 61-76.
Brym, Robert, Louise Birdsell Bauer and Mitch McIvor. (2013). “Is Industrial Unrest Reviving in
Canada? Strike Duration in the Early Twenty-First Century.” Canadian Review of Sociology.
50.2 (2013): 227-238.
Budd, John. “The Effect of Multinational Institutions on Strike Activity.” Industrial and Labor
Relations Review. 47.3 (1994): p. 401-416.
Budd, John. “Canadian Strike Replacement Legislation and Collective Bargaining: Lessons for
the United States.” Industrial Relations. 35.2 (1996): p. 245-260.
Budd, John. “The Effect of Strike Replacement Legislation on Employment.” Labour Economics.
7 (2000): p. 225-247.
Budd, John. and Yijiang. Wang. “Labor Policy and Investment: Evidence from Canada.”
Industrial and Labor Relations Review. 57.3 (2004): p. 386-401.
Cadillac Fairview Corporation v. R.W.D.S.U., [1989], 71 O.R. (2d) 206 (C.A.).
Campolieti, Michele, Robert Hebdon, and Benjamin Dachis. (2014). “The Impact of Collective
Bargaining Legislation on Strike Activity and Wage Settlements.” Industrial Relations. 53.3
(2014): p. 394-429.
Campolieti, Michele, Robert Hebdon, and Douglas Hyatt. “Strike Incidence and Strike Duration:
Some New Evidence from Ontario”. Industrial and Labor Relations Review. 58.4 (2005): p. 610630.
Carter, Donald. “The Duty to Bargain in Good Faith: Does it Affect the Content of Bargaining.”
Studies in Labour Law. Edited by Kenneth Swan and Katherine Swinton. Toronto: Butterworths,
1983, p. 35-54.
Chamberlain, Neil and James Kuhn. Collective Bargaining. New York: McGraw-Hill, 1965.
Cheung, Francis and Carl Davidson. “Bargaining Structure and Strike Activity.” The Canadian
Journal of Economics. 24.2 (1991): p. 345-371.
“Collective Bargaining Information.” Labour Program, 2015. Accessed on: 16 September 2015.
Retrieved from: http://www.labour.gc.ca/eng/resources/info/datas/wages/technical_notes.shtml.
Cousineau, J.M., R. Lacroix and D. Vachon. “Foreign Ownership and Strike Activity in Canada.”
Relations industrielles/Industrial Relations. 46.3 (1991): p. 616-631.
83
Collective Bargaining, Strikes and Lockouts
Cox, Archibald. “The Duty to Bargain in Good Faith,” Harvard Law Review. 71.8 (1958): p.
1401-1442.
Cramton, Peter, Morley Gunderson and Joseph Tracy. “Impacts of Strike Replacement Bans in
Canada.” Labor Law Journal. 50.3 (1999): p. 173-179.
Cramton, Peter, Morley Gunderson and Joseph Tracy. (1999). “The Effect of Collective
Bargaining Legislation on Strikes and Wages.” The Review of Economics and Statistics, 81.3
(1999): p. 475-487.
Curran, Bruce. “Event History Analysis of Grievance Arbitration in Ontario: Labour Justice
Delayed?” Paper Submitted as Portion of PhD Dissertation, University of Toronto, 2015.
Dachis, Benjamin and Robert Hebdon. “The Laws of Unintended Consequence: The Effect of
Labour Legislation on Wages and Strikes.” C.D. Howe Institute Commentary, 2010.
De Lint, Willem, Ryan Gostlow and Alan Hall. “Judgement by Deferral: The Interlocutary
Injunction in Labour Disputes Involving Picketing.” Canadian Journal of Law and Society. 20.2
(2005): p. 67-93.
Dinsdale, Henry and Dan Awrey. “Secondary Picketing in Canada: Thoughts for the Pepsi
Generation.” Queen’s Law Journal. 29.2 (2004): p. 789-808.
Doorey, David, “Could Manitoba’s Model on Work Stoppages Solve the NHL’s Problems?” Law
of Work (30 October 2012) online: <http://lawofwork.ca/?p=5800>.
Duffy, Paul and Susan Johnson. “The Impact of Anti-Temporary Replacement Legislation on
Work Stoppages: Empirical Evidence from Canada.” Canadian Public Policy. 35.1 (2009): p. 99120.
Eaton, Jonathan. “Is Picketing a Crime?” Relations industrielles/Industrial Relations. 47.1
(1992): p. 100-124.
Fang, Tony and Anil Verma. “Union Wage Premium.” Perspectives on Labour and Income. 14.4
(2002): p. 17-23.
Finley, Gregory. “Strike Lengths: Correcting for Prestrike Announcements and the Ratio of
Bargaining Size to Firm Size.” Journal of Labor Research. 31.4 (2010): p. 307-321.
Fudge, Judy. Fragmenting Work and Fragmenting Organizations: The Contract of Employment
and the Scope of Labour Regulation.” Osgoode Hall Law Journal. 44.4 (2007): p. 609-648.
84
Collective Bargaining, Strikes and Lockouts
Fudge, Judy and H. Glasbeek. (1995). “The Legacy of PC 1003.” Canadian Labour and
Employment Law Journal, 3, 357-400.
Galarneau, Diane and Thao Sohn. Long-term Trends in Unionization. Insights on Canadian
Society. Statistics Canada. 2013. Accessed on: September 17, 2015. Retrieved from:
http://www.statcan.gc.ca/pub/75-006-x/2013001/article/11878-eng.pdf
Godard, John. “Strikes as Collective Voice: A Behavioural Analysis of Strike Activity.” Industrial
and Labor Relations Review. 46.1 (1992): p. 161-175.
Godard, John. “Strikes and the Law.” Relations Industrielles/Industrial Relations. 53.2 (1998): p.
258-277.
Godard, John. “What Has Happened to Strikes?” British Journal of Industrial Relations. 49.2
(2011): p. 282-305.
Gomez, Rafael, and Danielle Lamb. "Union Wage Premiums in Canada Over Time"
(Unpublished Working Paper, 2015).
Gunderson, Morley, Robert Hebdon and Douglas Hyatt. “Strikes and Dispute Resolution”.
Labour-Management Relations in Canada, sixth edition. Edited by Morley Gunderson and
Daphne Taras. Toronto: Addison Wesley, 2009.
Gunderson, Morley, John Kervin and Frank Reid. “The Effect of Labour Relations Legislation on
Strike Incidence.” Canadian Journal of Economics. 22.4 (1989): p. 779-794.
Gunderson, Morley and Angelo Melino. “The Effects of Public Policy on Strike Duration.” Journal
of Labor Economics. 8.3 (1990): p. 295-316.
Harrison, Alan and Mark Stewart. “Cyclical Fluctuations in Strike Durations”. The American
Economic Review. 79.4 (1989): p. 827-841.
Haywood, Lee. “Replacement of Striking Workers During Work Stoppages in 1991”. Ontario
Ministry of Labour, Industrial Relations Division, Office of Collective bargaining Information, April
1, 1992.
Hebdon, Robert, Douglas Hyatt and Maurice. Mazerolle. “Implications of Small Bargaining Units
and Enterprise Unions on Bargaining Disputes: A Look into the Future?”. Relations
industrielles/Industrial Relations. 54.3 (1999): p. 503-526.
Jain, Harish and Parbudyal. Singh. “The Effects of the Use of Strike Replacements on Strike
Duration in Canada.” Labor Law Journal. 50.3 (1999): p. 180-186.
85
Collective Bargaining, Strikes and Lockouts
Kaufman, Bruce. “Research on Strike Models and Outcomes in the 1980’s: Accomplishments
and Shortcomings.” Research Frontiers in Industrial Relations and Human Resources. Edited by
David Lewin, Olivia Mitchell and Peter. Sherer. Madison, Wisconsin: Industrial Relations
Research Association, 1992, p. 77-129.
Kaufman, Bruce. “The Impossibility of a Perfectly Competitive Labor Market.” Cambridge
Journal of Economics. 31.5 (2007): p. 775-787.
Kaufman, Bruce. “The Theoretical Foundation of Industrial Relations and its Implications for
Labor Economics and Human Resource Management.” Industrial and Labor Relations Review.
64.1 (2010): p. 74-108.
Kaufman, Bruce and Michael Barry. “IR Theory Built on the Founders Principles with Empirical
Application to Australia.” ILR Review. 67.4 (2014): p. 1203-1234.
Kennan, John. “The Economics of Strikes.” Handbook of Labor Economics, third edition. Edited
by Orley Ashenfelter and David Card, 1986, p. 1091-1137.
Kochan, Thomas, Harry Katz and Robert McKersie. The Transformation of American Industrial
Relations. Ithaca, New York: Cornell University Press, 1986.
Labor Management Relations Act, 1975, 29 U.S.C.S., ss.151-67.
Langille, Brian. “Partnerships in Labour Law.” Osgoode Hall Law Journal. 21.3 (1983): p. 496536.
Langille, Brian. “Has Ontario’s Anti-Scab Law Made any Difference – Legally, Economically, or
Otherwise?” Canadian Labour & Employment Law Journal. 3 (1995): p. 461-474.
Langille, Brian and Patrick Macklem. “Beyond Belief: Labour Law’s Duty to Bargain.” Queen’s
Law Journal. 13.1 (1988): p. 62-102.
Legree, Scott, Tammy Schirle and Mikal Skuterud. “The Effects of Labour Relations Laws on
Union Density Rates: Evidence from Canadian Provinces.” Working Paper no. 141. Canadian
Labour Market and Skills Researcher Network, 2014.
MacDonald, Darren. “Minister won’t ‘close the door’ on replacement worker ban.” Northern Life
24 July 2015. Accessed on: 16 September 2015. Retrieved from:
www.northernlife.ca/news/localNews/2015/07/24-labour-sudbury.aspx.
MacNeil, Michael. “Labour Picketing and Consumer Boycotts: Judicial Ideology in Kmart and
Allsco.” Canadian Labour and Employment Law Journal. 8 (2000): p. 79-94.
86
Collective Bargaining, Strikes and Lockouts
Macklem, Patrick. “Secondary Picketing, Consumer Boycotts and the Charter.” Canadian
Labour and Employment Law Journal. 8 (2000): p. 1-15.
Manitoba, Labour Relations Amendment Act, 2000, (2), S.M. 2000, c.45.
Manitoba, Labour Relations Amendment Act, 2000, S.M. 2004, c.31.
Martin, Peter. “Global Business: Prepare for Strikes.” CPM Global Assurance, 5.5, May 2008.
Martinello, Felice and Ronald Meng. “Effects of Labor Legislation and Industry Characteristics
on Union Coverage in Canada.” Industrial and Labor Relations Review. 46.1 (1992): p. 176-190.
Mojtehedzadeh, Sara. “Metro deal sets new precedent for fair wages and schedules”, The
Toronto Star, July 17, 2015.
National Labor Relations Act, 1935, s.8(d).
National Labor Relations Board v. Fruit and Vegetable Packers & Warehousemen, Local 760 et
al., [1964] 377 U.S. 58.
National Labor Relations Board v. MacKay Radio and Telegraph Co., [1938] 304 U.S. 333,346.
Ontario, Labour Relations Act, 1948, S.O. 1948, s.12.
Ontario, Labour Relations Act, 1975, S.O. 1975, c.76, s. 21(1).
Marco Chown Oved, “Crown factory workers’ strike ignored for 17 long months,” The Toronto
Star (2 February 2015) online: <www.thestar.com/news/gta/2015/02/02/crown-factory-w
orkers-strike-ignored-for-17-long-months.html>.
Palmer, E. E. “The Myth of “Good Faith” in Collective Bargaining.” Alberta Law Review. (19651966): p. 409-418.
Palys, Ted. Research Decisions: Quantitative and Qualitative Perspectives, 3rd ed. (Toronto:
Thomson Nelson, 2003).
Parry Sound (District) Social Services Administration Board v. Ontario Public Services
Employees Union, Local 324, [2003] 2 S.C.R. 157.
Peters, John. “Down in the Vale: Corporate Globalization, Unions on the Defensive, and the
USW Local 6500 Strike In Sudbury, 2009-2010.” Labour/Le Travail. 66 (2010): p. 73-105.
Pink, Ronald and David Wallbridge. “The Future of Labour Arbitration.” Ottawa: Presentation at
The 2010 Administrative, Labour and Employment and Privacy And Access Law Conference,
November 26-27, 2010.
87
Collective Bargaining, Strikes and Lockouts
Ponak, Allen and Corliss Olson. “Time Delays in Grievance Arbitration in Alberta.” Relations
industrielles/Industrial Relations. 47.4 (1992): p.690-708.
Ponak, Allen, Wifred Zerbe, Sarah Rose and Corliss Olson. “Using Event History Analysis to
Model Delay in Grievance Arbitration.” Industrial and Labor Relations Review. 50.1 (1996): p.
105-121.
R. v. Layton, [1986] 33 C.C.C. (3d) 550 (Ont. Prov. Ct.).
Rees, Albert. “Industrial Conflict and Business Fluctuations.” Industrial Conflict. Edited by Arthur
Kornhauser, Robert Dubin, and Arthur Ross. New York: McGraw-Hill, 1954, p. 213-221.
Rose, Joseph. “Statutory Expedited Grievance Arbitration: The Case of Ontario.” The Arbitration
Journal. 41.4 (1986), p. 30-45.
Sack, Jeffrey, Emma Phillips and Hugo Leal-Neri. “Protecting Workers in a Changing
Workworld: The Growth of Precarious Employment in Canada, the United States and Mexico.”
The Employment Relationship: A Comparative Overview. Edited by Guiseppe Casale. Oxford:
Hart, 2010, p. 233-285.
Sangster, Joan. “’We No Longer Respect the Law’: The Tilco Strike, Labour Injunctions and the
State.” Labour/Le Travail. 53 (Spring 2004): p. 47-87.
Savage, Larry and Jonah Butovsky. “A Federal Anti-Scab Law for Canada? The Debate over
Bill C-257.” Just Labour: A Canadian Journal of Work and Society. 13 (2009): p.15-28.
“Employment Insurance Regular Benefits.” Service Canada. 2015. Accessed on: 16 September
2015. Retrieved from: http://www.servicecanada.gc.ca/eng/ei/types/regular.shtml.
Shilton, Elizabeth. “Choice, but No Choice: Adjudicating Human Rights Claims in Unionized
Workplaces in Canada.” Queen’s Law Journal. 38.2 (2013): p. 461-510.
Shniad, Sid. “The Decline of Organized Labour and How its Current Trajectory Can Be
Reversed.” Labour/Le Travail. 65 (2010): p. 141-148.
Singh, Parbudyal, Deborah Zinni, and Harish Jain. “The Effects of the Use of Striker
Replacement Workers in Canada: An Analysis of Four Cases.” Labor Studies Journal. 30.2
(2005): p. 61-85.
Smith, Charles. ““We Didn’t Want to Totally Break the Law”: Industrial Legality, the Pepsi Strike,
and Workers Collective Rights in Canada.” Labour/Le Travail. 74 (Fall 2014): p. 89-121.
88
Collective Bargaining, Strikes and Lockouts
Smith, Douglas. “The Determinants of Strike Activity in Canada.” Relations
industrielles/Industrial Relations. 27.4 (1972): p. 663.
Stanford, Jim. “Management has Warmed to Work Stoppages.” The Globe and Mail 20 January
2012. Accessed on: 21 September 2015.
Thornicroft, Kenneth. “Sources of Delay in Grievance Arbitration.” Employee Responsibilities
and Rights Journal. 8.1 (1995): p. 57-66.
Thornicroft, Kenneth. “The grievance arbitration process and workplace conflict resolution”, In
M. Gunderson, & D. G. Taras (Eds.), Canadian labour and employment relations (6th ed.).
Toronto: Pearson Addison Wesley.
Tucker, Eric. “Hersees of Woodstock Ltd. V. Goldstein: A Small Town Case Makes it Big.” Work
on Trial: Canadian Labour Law Struggles. Edited by Judy Fudge and Eric Tucker. Toronto:
Irwin, 2010, p. 217-248.
Verma, Anil, Jeffrey G. Reitz, and Rupa. Banerjee. “Unions, Race, Immigrants and Earnings: A
Longitudinal Examination of the Effect of Union Membership on the Income Progression of
Immigrants to Canada.” International Migration Review. (Forthcoming).
Vosko, Leah. Temporary Work: The Gendered Rise of a Precarious Employment Relationship.
Toronto: University of Toronto Press, 2000.
Vosko, Leah. “A New Approach to Regulating Temporary Agency Work in Ontario or Back to the
Future?”. Relations industrielles/Industrial Relations. 65.4 (2010): p. 632-653.
Vosko, Leah. Managing the Margins: Gender, Citizenship and the International Regulation of
Precarious Employment (New York: Oxford, 2010).
Walsworth, Scott and Richard Long. “Is the Union Employment Suppression Effect Diminishing?
Further Evidence from Canada.” Relations Industrielles/Industrial Relations, 67.4 (2012): p. 654680.
Weil, David. The Fissured Workplace: Why Work Became So Bad For So Many and What Can
Be Done to Improve It. (Cambridge: Harvard University Press, 2014).
Whitaker, Kevin. “The Development and Use of Mediation: Arbitration in Ontario.” Arbitration
2009: Due Process in the Workplace: Proceedings of the 62nd Annual Meeting, National
Academy of Arbitrators. Edited by Paul D. Staudohar. Arlington, VA: BNA Books, 2010, p. 209214.
Winkler, Warren, (Hon). “Labour Arbitration and Conflict Resolution: Back to Our Roots.”
Queen’s University. Kingston, ON. 30 November 2010. Donald Wood Lecture.
89
Collective Bargaining, Strikes and Lockouts
Winkler, Warren, (Hon.). “Arbitration as Cornerstone of Industrial Justice.” Speech presented at
Queen’s University, Kingston, ON. 30 November 2011. Available at:
http://www.ontariocourts.ca/coa/en/ps/speeches/2011-arbitration-cornerstone-industrialjustice.htm
Zatz, Noah. “’Who is an Employee?’ and Other Questions”. Prepared for the NLRB/George
Washington University Symposium on “The National Labor Relations Act at 75.” It’s Legacy and
its Future.” 2010. Manuscript available at http://www.nlrb.gov/75th/Documents/Zatz.pdf.
90
Collective Bargaining, Strikes and Lockouts
Appendix 1 - Additional Graphs of Work Stoppages Data
Figure A1 – Aggregate Annual Number of Workers Involved in Work Stoppages,
Strikes, and Lockouts in Canada, 1980-2014.
Figure A2 - Aggregate Number of Workers Involved in Work Stoppages, Strikes
and Lockouts as Ratio of Total Unionized Employees (Canada) (all unit sizes)
91
Collective Bargaining, Strikes and Lockouts
Figure A3 – Average Duration of Work Stoppages , Strikes and Lockouts in
Canada, 1980-2014 (all unit sizes)
92
Collective Bargaining, Strikes and Lockouts
Appendix 2 - Summary of Research Design (Jurisdiction, Timeframe, and Data) in
Quantitative Studies of The Effects of Specific Components of Labour Policy on
Strike Dynamics and Other Variables
Author (year)
Jurisdiction & Timeframe Studied
Gunderson, Morley,
Canada, all jurisdictions, bargaining units of either
Kervin, John and Reid,
200+ or 500+ employees,
Frank (1989)
Unit-level data
1967-1985
Gunderson, Morley and
Canada, all jurisdictions, bargaining units of either
Melino, Angelo (1990)
200+ or 500+ employees
Unit-level data
1967-1985
Budd, John (1996)
Canada all jurisdictions,
Unit-level data (manufacturing sector only)
1966-1985
Cramton, Peter,
Canada, all jurisdictions
Gunderson, Morley and
Unit-level data, large (500+ employees) units only
Tracy, Joseph (1999)
1967-1993
Budd, John (2000)
Canada, all provinces
Monthly, provincial aggregate data
1966 – 1994
Budd, John and Wang,
Canada, all provinces,
Yijiang (2004)
Monthly, provincial aggregate data
1967-1999
Campolieti, Michele,
Ontario,
Hebdon, Robert and
Unit-level data
Hyatt, Douglas (2005)
1984-1992
Duffy, Paul and Johnson,
Canada, all jurisdictions
Susan (2009)
Monthly, provincial aggregate data
1978-2003
93
Collective Bargaining, Strikes and Lockouts
Dachis, Benjamin and
Canada, all jurisdictions
Hebdon, Robert (2010)
Unit-level data
1978-2008
Campolieti, Michele,
Canada, all jurisdictions,
Robert Hebdon, and
Unit-level data
Benjamin Dachis (2014)
1978-2008
Legree, Scott, Schirle,
Canada, all jurisdictions
Tammy and Skuterud,
Monthly, provincial aggregate data
Mikal (2014)
1981-2012
94
Collective Bargaining, Strikes and Lockouts
Appendix 3 – Provisions in Manitoba’s Labour Relations Act, C.C.S.M. c. L10,
Pertaining to Subsequent Interest Arbitration in Response to Protracted Work
Stoppages (Initial Provisions Adopted in 2000, and Amended in 2004).
SETTLEMENT OF SUBSEQUENT AGREEMENTS
Dispute about subsequent agreements
87.1(1) Where a collective agreement has expired and a strike or lockout has commenced, the employer or the
bargaining agent for a unit may apply in writing to the board to settle the provisions of a collective agreement if
(a) at least 60 days have elapsed since the strike or lockout commenced;
(b) the parties have attempted to conclude a new collective agreement with the assistance of a conciliation
officer or mediator for at least 30 days during the period of the strike or lockout; and
(c) the parties have not concluded a new collective agreement.
Notice
87.1(2)
The board shall promptly notify the parties when it receives an application.
Board to determine if good faith bargaining
87.1(3)
On receiving an application, the board shall inquire into negotiations between the parties and determine
(a) whether or not they are bargaining in good faith in accordance with subsection 63(1); and
(b) whether or not they are likely to conclude a collective agreement within 30 days if they continue bargaining.
Determination within 21 days
87.1(3.1) Except in the circumstance mentioned in subsection (4), the board shall make its determination under
subsection (3) within 21 days after it has notified the parties of the application, even if an unfair labour practice
complaint has been filed under subsection 30(1) alleging a failure to bargain in good faith under subsection 63(1).
Discretion of board
87.1(4) The board may delay making a determination under subsection (3) until it is satisfied that the party
making the application has bargained sufficiently and seriously with respect to those provisions of the collective
agreement that are in dispute between the parties.
S.M. 2000, c. 45, s. 23; S.M. 2004, c. 31, s. 2.
No settlement if good faith bargaining and agreement is likely
87.2(1) If the board finds under subsection 87.1(3) that the parties are bargaining in good faith and are likely to
conclude a collective agreement within 30 days if they continue bargaining, it shall decline to settle the provisions
of a collective agreement between them and notify them of that fact. The board may, however, appoint a board
representative, or request the minister to appoint a conciliation officer, to confer with the parties to assist them in
settling the provisions of a collective agreement.
New application if no agreement within further 30 days
87.2(2) If 30 days have elapsed since notice was given under subsection (1) and the parties have failed to
conclude a collective agreement, either party may make a new application to the board under subsection 87.1(1).
S.M. 2000, c. 45, s. 23.
Settlement
95
Collective Bargaining, Strikes and Lockouts
87.3(1) If the board determines under subsection 87.1(3) that the party making an application under
subsection 87.1(1) is bargaining in good faith but that a new collective agreement is unlikely to be concluded
within 30 days if the parties continue to bargain,
(a) the employees shall immediately terminate any strike;
(a.1) the employer shall immediately terminate any lockout;
(b) the employer shall reinstate the employees as provided for in subsection 87(5); and
(c) the provisions of a collective agreement between the parties shall be settled
(i) by an arbitrator, if the parties serve a notice of their wish for arbitration under subsection (2), or
(ii) by the board within 90 days of its finding, in any other case.
New application
87.3(1.1) If the board determines under subsection 87.1(3) that the applicant party is not bargaining in good faith,
that party may at any time after the determination is made make a new application under subsection 87.1(1) for
the board to settle the provisons of a collective agreement.
Arbitration
87.3(2) Within 10 days after a determination by the board that the applicant party is bargaining in good faith but
that a new collective agreement is unlikely to be concluded through further bargaining, the employer and the
bargaining agent may serve a notice on the board stating that they wish to have the collective agreement settled
by arbitration. The notice must name a person who has agreed to act as arbitrator.
Arbitrator to settle collective agreement
87.3(3) The arbitrator shall settle the provisions of the collective agreement within 60 days after notice is served
on the board under subsection (2).
Arbitration provisions of this Act apply
87.3(4) The provisions of this Act respecting arbitration apply, with necessary modifications, to an arbitrator
acting under this section.
Term of collective agreement
87.3(5) Subject to subsection (5.1), a collective agreement settled by an arbitrator or the board under this
section is effective for a period of one year following the expiry date of the previous collective agreement, or for
any longer period the parties agree to.
Extension of term of agreement
87.3(5.1) A collective agreement settled by an arbitrator or the board more than six months following the expiry
date of the previous collective agreement shall remain in effect for six months following the date of settlement.
Collective agreement binding
87.3(6) A collective agreement settled under this section is binding on the parties and on the employees in the
unit as though it were a collective agreement voluntarily entered into between the parties, but the parties may
nevertheless amend its provisions by a subsequent written agreement.
Subsections 87(6) and (8) apply
87.3(7) Subsections 87(6) and (8) apply, with necessary changes, to the settlement of a collective agreement
under this section.
S.M. 2000, c. 45, s. 23; S.M. 2004, c. 31, s. 3.
Review
96
Collective Bargaining, Strikes and Lockouts
87.4
The minister shall request the Manitoba Labour Management Review Committee to review the operation
of sections 87.1 to 87.3 at least once in each 24-month period after those sections come into force and provide a
report to the minister setting out their findings. The minister shall table the report in the Legislative Assembly as
soon as possible after receiving it.
S.M. 2000, c. 45, s. 23.